Professional Documents
Culture Documents
Editorial Board
David F. Batten
Manfred M. Fischer
Geoffrey J. D. Hewings
Peter Nijkamp
Folke Snickars (Coordinating Editor)
With 77 Figures
and 27 Tables
Springer
Prof. Dr. DONALD G. JANELLE
University of California, Santa Barbara
Center for Spatially Integrated Social Science
Santa Barbara, CA 93106-4060
USA
The use of general descriptive names, registered names, trademarks, etc. in this publication does not
imply, even in the absence of a specific statement, that such names are exempt from the relevant protec-
tive laws and regulations and therefore free for general use.
Cover design: Erich Kirchner, Heidelberg
Typesetting: Camera ready by editors
SPIN 10733972 4212202/xz Printed on acid-free paper 543 2 I 0
Foreword
The use of the term the information age to describe the period that we now fmd
ourselves living in is open to misinterpretation. Society has always been based on
exchanging information, and our libraries have long been rieh sources of vast
quantities of readily available information; it is information technologies that have
changed rapidly sinee the invention of the digital computer. These technologies
are themselves products of long-term societal processes: The eeonomic desire to
shorten the time that lapses between produetion and consumption of eommodities,
annihilating space with time; the political desire to control such large-scale sys-
tems as commodity ehains, nations, and the military; and the human desire to lib-
erate ourselves from the constraints of our loeal daily lives. They also have had
profound effeets on societal proeesses. One of the most widely discussed effeets,
and a eonsistent theme of this volume, is that the information age is bringing about
the end of geographie al distance as a signifieant baITier ofhuman interaction.
This claim underlies prognostications about the information age: That this will
be the age of globalization; of the global village; of the liberation of human inter-
action from the tyranny of space; of the dissolution of cities and workplaces; of
the plugged-in soeiety; and of the surveillance society. If these prognostications
were true, then the topie of aceessibility would indeed be a disappearing research
pro gram and this book a marker of its disappearance. Yet, things are much more
complicated than this; the demise of distanee has been greatly exaggerated. While
there is a germ of truth to these prognostications, as there must be for them to
resonate as they do, they often disguise more than they reveal. Flows of informa-
tion are possible almost irnmediately over distances of arbitrary length, but this
does not mean that everyone is equally accessible to everyone else. Rather, the
geometry of the information age approximates the hypothesized worrnholes of
quantum physics - instantaneous connections between those who are plugged in to
the right equipment, while neighbors remain off-line and inaceessible. Geographie
and non-geographie information are available in unpreeedented quantities, but
they aceumulate in the hands of certain social actors whereas others are excluded
- creating black holes where information seems to disappear from social view.
Even the ultimate distance-Iess society, cyberspace, becomes un-navigable with-
out using spatial metaphors to make sense of it, and is eonnected in complex but
predictable ways to the differentiated material spaees of society.
This eolleetion of essays takes up the challenge of rethinking what aeeessibility
means and how to measure it in the information age, with particular attention to
geographie information. It addresses not only aecessibility between those who are
plugged in, and the geography of cyberspace, but also differences in aecessibility
to information technologies and the relationship between cyber-accessibility and
aecessibility on the ground. In doing so, the authors revive what has been an im-
vi Foreword
portant but theoretically moribund concept; breathing new life into the concept of
accessibility, and challenging preconceptions about its demise. They also move
beyond attempts to equate accessibility with an exogenous Newtonian metric of
Euclidean distance to unpack how accessibility is a construct of socia1 practices.
The conversations that lie behind this book were catalyzed and made possible by
a conference organized under the auspices of the National Center for Geographie
Information and Analysis. This conference, Measuring and Representing Accessi-
bility in the Information Age, was held in November 1998, at the Asilomar Con-
ference Center in Pacific Grove, Califomia. It was one of aseries of nine meetings
organized by NCGIA between 1997 and 1999 to advance the research agenda of
geographie information science, under the Varenius Project (funded by the Na-
tional Science Foundation, NSF Grant SBR-9600465). These nine meetings were
equally divided among three areas of focus: Geographies of the Information Soci-
ety; Cognitive Models of Geographie Space; and Computational Implementations
of Geographie Concepts (http://www.ncgia.ucsb.edu:80/varenius). The accessibil-
ity meeting was held within this first area. The Geography of the Information So-
ciety Panel, chaired by Eric Sheppard and including Helen Couclelis, University
of Califomia at Santa Barbara, Stephen Graham, Newcastle University, UK, lW.
Harrington, Jr., University of Washington, and Harlan Onsrud, University of
Maine at Orono, also organized meetings on Place and Identity in an Age ofTech-
nologically Regulated Movement, and Empowerment, Marginalization and Public
Participation GIS. The Panel conceived the topic of measuring and representing
accessibility, but the success of this meeting was due to the efforts of David
Hodge and Donald Janelle in bringing the idea to fruition. Under their exceptional
organizational skills, together with those ofNCGIA staff LaNell Lucius and Abby
Caschetta, a stimulating three-day meeting occurred at which preliminary vers ions
of the chapters that follow were presented. This book is exemplary of how the
Varenius Project is catalyzing and making available new research within areas
central to geographie information science.
1 Direetor ofProjeet Varenius and Chair ofthe Exeeutive Committee ofthe National Center
for Geographie Information and Analysis. Department of Geography, University of
Califomia, Santa Barbara CA 93106-4060, USA. Email: good@negia.uesb.edu
2 Chair ofthe Varenius Panel on Geographies ofthe Information Soeiety. Department of
Geography, University ofMinnesota, Minneapolis MN 55455, USA.
Email: sheppOO l@maroon.te.umn.edu
Preface
The objectives of this book are to broaden understanding of conceptual and ana-
lytical approaches for accessibility research appropriate to the information age,
and to demonstrate possible contributions for geographic information science in
representing the geographies of the information society. In seeking to meet these
objectives, the editors and authors highlight significant linkages among informa-
tion resources, traditional places, and cyberspace, and focus on expanding models
of space (and time) that encompass both the physical and virtual worlds.
The origins of this book stern from two multi-disciplinary conferences spon-
sored by the National Center for Geographic Information and Analysis (NCGIA).
The first was the September 1996 conference in Baltimore on Spatial Teehnolo-
gies, Geographie Information and the City. The second, from which the chapters
of this book originate, was the November 1998 conference at the Asilomar Con-
ference Center in Pacific Grove, Califomia on Measuring and Representing Ae-
eessibility in the Information Age. This book is structured around the primary
themes of that meeting. Part I explores the conceptualization and measurement of
accessibility; Part II focuses on the visualization and representation of information
space within Geographic Information Systems (GIS) and other computerized dis-
play systems, and Part III considers the social issues that should inform the meas-
urement and representation of accessibility. Each of these parts is preceded with
an integrative essay that links the individual chapters to the broader literature on
accessibility - primarily from geography, regional science, and planning. In Chap-
ter 1, the Editors offer an explanation for the book's title, casting a wide perspec-
tive that focuses on the resource role of information, the importance of
accessibility in the everyday life of places, and the co-adaptation of societal struc-
tures and cyberspace.
Special recognition is given to Helen Couclelis, who inspired the proposal for a
Varenius initiative on accessibility in the information age. She organized the con-
ference in Baltimore and was instrumental in placing accessibility on the agenda
ofthe Varenius project. In Part IV, the Conclusion, she broadens the scope ofthis
collection, raising issues regarding the sustainability of current societal accessibil-
ity practices in the interrelated realms of transportation and communication.
We thank those who made this book possible, beginning with Michael Good-
child (Director of the NCGIA's Varenius Project); he orchestrated the pre-
conditions for sponsoring a broad range of research and conference initiatives.
Members of the Varenius Panel on Geographies of the Information Society ac-
cepted a proposal to foster research on issues relating to accessibility, and we owe
special thanks to the Chair of the Panel, Eric Sheppard, for support and advice at
all stages of this project. The Steering Committee for organizing the conference at
Asilomar - Michael Batty, Helen Couclelis, Arthur Getis, Harvey Miller, and
viii Preface
1 Center for Spatially Integrated Social Science, University of Califomia, Santa Barbara CA
93106-4060, USA. Email: janelle@ncgia.ucsb.edu
2 College of Arts and Sciences, University ofWashington, Seattle WA 98195-3765, USA.
Email: hodge@u.washington.edu
Contents
Foreword v
MICHAEL F. GOODCHILD AND ERlC SHEPPARD
Preface Vll
Introduction
Figures 357
Tables 361
Contributors 377
Introduction
1 Information, Place, Cyberspace, and
Accessibility
1 Center for Spatially Integrated Social Science, University of Califomia, Santa Barbara CA
93106-4060, USA. Email: janelle@ncgia.ucsb.edu
2 College of Arts and Sciences, University ofWashington, Seattle WA 98195-3765, USA.
Email: hodge@u.washington.edu
1.1 Introduction
This chapter provides a broad overview of alternate ways for seeing the operative
linkages between human experiences on the ground (in pi ace) and user experi-
ences in cyberspace. Information is treated as the resource that binds these realms
into functional human systems, while computer, telecommunication, and transpor-
tation technologies are viewed as tools of accessibility that are allocated dift'eren-
tially among people, institutions, and regions. Two general propositions guide the
discussion. First, there are significant structural linkages among information re-
sources, traditional places, and cyberspace; and second, grasping these linkages
requires expanded models of space (and time) that encompass both the physical
and virtual worlds.
The operational definitions of 'information', 'place', 'cyberspace', and 'accessibil-
ity' are seen as open-ended, subject to new interpretation in different social situa-
tions and technical domains. Information lies on a continuum ±rom raw facts to
knowledge and can be see as the outcome of creative manipulation of data and
previous insight to assist our arrival at new thresholds of understanding. In the
context of this book, information is one of the critical resources of the new econ-
omy. Information figures centrally in the processes for producing and allocating
goods and services.
Place is seen as an extended locale of human activity imbued with the heritage,
identity, and commitment ofpeople and institutions. It is often linked with notions
of family, neighborhood, and community (e.g., my hometown). However, the dis-
tinctive qualities of some places extend their recognition far beyond the local do-
main of immediate and long-term experience. Thus, Manhattan is a place in the
minds of those who have never been there and for those who share its ambiance in
a more transitory state (e.g., visitors). As with information, the meaning ofplace is
subject to transformation through social and technological innovation, and through
various levels and means of association and experience.
4 D.G. Janelle and D.C. Hodge
1 Helen Couclelis, ed. 1996. Spatial Teehnologies, Geographie Information, and the City,
Teehnieal Report 96-10 (Deeember). (Santa Barbara: National Center for Geographie In-
formation and Analysis), p. 6.
Information, Place, Cyberspace, and Accessibility 5
notions of distance and connectivity that are insufficient for understanding new
forms of structures and behaviors characterizing an information age. Accessibility
and spatial interaction in the traditional physical sense remain important, but in-
formation technologies are dramatically moditying and expanding the scope of
these core geographical concepts. Through technological, structural, and social
developments, an increasing range of transactions takes place in virtual space, or
in some new hybrid space combining the physical with the virtual. Of importance
also is the influence of new forms of communication on the use of and investment
in traditional transportation infrastructure. Moreover, just as space can be frag-
mented so too can time, as activity rhythms in one place become increasingly syn-
chronized with those in distant places. Geographic information science and
technology, themselves products of this new information age, potentially have a
major role to play in helping to reconceptualize, measure, represent, monitor, and
plan for the new emergent geographies of accessibility.
Goodchild and Sheppard (see Foreword) recognize correctly that society has al-
ways depended on information and has gone to great lengths in promoting signifi-
cant information resources - from the notable centralized repositories at
Alexandria to the Library of Congress. However, the nineteenth-century move-
ment for public libraries in every city and town and the recent emergence of a
globally interconnected Internet are models of information resources that tran-
scend many of the barriers to building informed, knowledge-empowered societies.
Hence, it is not information alone that distinguishes the early twenty-first century
Information Age; rather, the Information Age is distinguished by the technologies
for disseminating information and by the growing dominance of information in-
dustries over the economy at large. Batty and Miller (Chapter 8) observe that in-
formation is replacing energy as the basis for organizing economies and societies.
The power and freedom that information offers motivate the quest for technolo-
gies that provide nearly instant connectivity among institutions and individuals.
The reality of technology's promise to annihilate the effects of distance on hu-
man commerce and discourse falls short of universal application. Societal proc-
esses that allocate the tools and knowledge to make this happen are embedded in
legacies of differential access to infrastructure that are difficult to overcome. Mar-
ket forces allocate information differentially over space and selectively among
people and institutions, and the commodification of information poses a price bar-
rier to access. Constrained by limits on financial resources and knowledge, many
people and institutions are bounded by systems of social and political discrimina-
tion.
Information technology provides the potential to make all places equally acces-
sible to opportunity. However, society's allocation ofthe tools for overcoming the
6 D.G. Janelle and D.C. Hodge
intrinsic baITiers of distance may feed growth in the relative inequality among
places, thus increasing rather than diminishing the underlying complexity to the
human geography of accessibility. Assessment of such a proposition makes the
measurement and representation of accessibility important scientifically and sig-
nificant socially.
It is important to assess critically the assumption that equal and unlimited acces-
sibility is necessarily in the long-term best interest of society. For instance, for
access to information, quantity may be less important than quality and timeliness.
Sui (Chapter 7) sees the inundation of information posing threatening prospects
for information-overloaded and dysfunctional societies. Separating out the crea-
tive use of information in ways that enhance individual accomplishment and
societal benefit may not be easy. Thus, ways to filter information may be as, or
more, important to some than accessibility. In the realm of cyberspace, as Dodge
(Chapter 11) observes, the ratio ofnoise to information may be very high.
Other attributes of information are also important and need to be factored into
our discussion. Hanson (Chapter 16) speaks of 'collective information assets' and
maintains that diversity of information sources enhance the choices and range of
opportunities within communities. The social context of information access also
derives from constitutional rights and legislative initiatives that can shift the con-
trol of information among public and private sector agents (see Onsrud, Chapter
18). These many facets of information resources suggest that there is no single,
easy way to measure and represent accessibility.
The roles of places as repositories and conveyors of social capital are central to
any consideration of accessibility. In Chapter 16, Hanson sees place as a funda-
mental construct that needs to be strengthened, possibly through selective use of
communication technologies that enhance community building and human net-
working capabilities. She sees jobs as important stabilizing linkages to place for
most people. Scott (Chapter 3) and Shen (Chapter 4) make explicit attempts to
differentiate job accessibility at metropolitan levels. While transportation is the
operative tool for achieving job access, they recognize that telecommunications
and Internet access represent a growing factor in the ability to match employees
withjob opportunities.
Adam's (Chapter 13) offers a highly explicit representation ofhuman travel and
communication behavior. His use of activity diaries and CAD-based visualizations
illustrate how individual daily activities are embedded in both virtual and place-
based networks. Extensibility linkages (the ability of people to engage with distant
locations) yield virtual presence or participation beyond the local realm, but the
bounded nature of everyday life at the local level remains a paramount factor in
the lives of all subjects in his investigation. Kwan reinforces this observation in
Information, Place, Cyberspace, and Accessibility 7
Although they clearly have the capability of complementing one another, place
and virtual space share an uneasy alliance. In part, this is because of the recent,
rapid, and potentially destabilizing consequences of cyberspace. However, this
uneasy relationship mayaiso reflect insufficient attention to analyzing the struc-
ture of cyberspace and its patterns of use. Uncovering the structure of cyberspace
poses problems in the use oftraditional concepts ofmorphology and distance. The
physical hardware has the character of traditional infrastructure networks - paths,
switching centers, relays, and capacity limits. But the process of use is invisible
and does not conform to any strict Newtonian metric. The lack of a centralized
system to monitor flows of activity over the network makes geographical interpre-
tation difficult. Thus, Batty and Miller (Chapter 8) express concern for vagueness,
fluidity, and low accountability for either the content or the quality of interactions
on the Internet.
The scarcity of reliable quantitative indicators for analysis may be a major in-
hibitor to research on the nature of cyberspace. However, this book offers a few
important examples of different approaches. Moss and Townsend (Chapter 10)
8 D.G. Janelle and D.C. Hodge
Separating out the lack of interaction independently from the lack of access is
problematic in any effort to measure and represent accessibility. Inherited urban
structure, for example, poses a significant constraint and should be considered in
any attempt to assess the real impact of virtual technologies on future structures.
With the right hardware and software tools at their disposal, individuals now have
the capacity to selectively turn on and off their engagement with the world of in-
formation, but there is significantly less flexibility in their ability to escape the
structural confines of their immediate physical environments. This book illustrates
theoretical approaches for dealing with such issues, seen for example in Heikkila's
development of a fuzzy-logic framework for understanding the complexity of ac-
cessibility (Chapter 6). In addition, Adams (Chapter 13) and Kwan (Chapter 14)
illustrate the uses of 3-D CAD and GIS representations to depict how space-time
diaries help capture the dynamics of behavior at the individual level. These au-
thors do not claim final solutions to the problems of representation, but they do
demonstrate that new space-time typologies, and new topologies, are needed to
accommodate the interdependence of both the physical and virtual worlds of eve-
ryday life.
A hybrid blend of physical and virtual space may now constitute the new geog-
raphy of the information age; and, as Batty and Miller note (Chapter 8), it may
rein force patterns of physical infrastructure that impinge on the comparative levels
of physical accessibility for different regions. For example, on-line book pur-
chases still require transshipment facilities to accommodate the physical move-
ment from the producer to the consumer. Thus, it is interesting to consider how
virtual spaces map onto traditional conceptions of geographic space, and vice
versa, and to address the issue of how we handle such complexities analytically.
For example, how can traditional spatial interaction and spatial gradients within
hybrid spaces (and space-times) be visualized with GIS?
Are there information counterparts to the accessibility and potential surfaces
developed by regional scientists for interaction in physical space? Scott (Chapter
3) considers the issue of scale and of potential versus realized accessibility in
looking at job access in the Los Angeles region, and Shen (Chapter 4) has devised
indexes and composite measures of access to consider both commuting and tele-
commuting options within the Boston metropolitan area.
Data availability is a constraint on advances in the area of accessibility research.
The traditional national census fails to capture information on telecommunication
and Internet activity, and correspondingly omits possibilities to ac count for the
dynamics of cross-border commerce or for individual lifestyle changes in response
to the potentials of information-age technologies. Survey designs illustrated here
by Adams (Chapter 13), and Kwan, (Chapter 14), and simulation methodologies
suggested by Forer and Huisman (Chapter 5), extend the scope oftime-geography
framework for depicting human space-time behavior. And, as illustrated by Har-
vey and Macnab (Chapter 9), real-time accessibility for face-to-face or interactive
10 D.G. Janelle and D.C. Hodge
1. 7 Conclusion
Part III present critiques of methodologies that fail to account for broad social
policy and theoretical perspectives.
Sui raises the possibility that society has changed beyond the point of relevance
for worrying too much about accessibility measurement. He sees greater validity
in a more evolutionary pattern of change that is best characterized as 'adaptation'
instead of accessibility. Hanson (Chapter 16) argues that there are 'silenees' in our
existing measures of accessibility, omission of which could foster narrowness of
interpretations regarding the fundamental importance of place-based networks of
people and institutions. Indeed, Wilson (Chapter 15) reminds us that the opportu-
nity costs of accessibility should not be neglected. Mugerauer (Chapter 19) ex-
tends this argument, presenting a forceful case for empowerment in the use of
measurement and descriptive tools. He raises concerns about the culturally ho-
mogenizing influences of common information pools and analytic methodologies.
He questions whether or not current spatial analytic tools, with built-in assump-
tions about the nature of space, such as GIS, can be adapted to account for local
interests and cultural perspectives.
In the concluding section, Part IV, Couclelis adopts a policy perspective that
focuses on the potential importance of accessibility analysis to inform the trans-
formation of spatial structures in ways that are more sensitive to environmental
and societal cost constraints. She argues that automobile dependence might be
lowered through new ways of uniting transportation and communication technolo-
gies. This view places the accessibility theme in a central position for the transpor-
tation and communication planning of cities and regions. She observes how
transport uses of information technologies offer a new order of flexibility for
automobile users and a new order of uncertainty for planners. Nonetheless, she
does not foresee an inevitable outcome of deteriorating, uncontrolled environ-
mental change. She reminds us that societies can shape these technologies and
their uses to achieve sustainable environments. Similarly, Occelli (Chapter 17)
sees the policy uses of informed measures and models as meriting a high priority
in the research agenda of regional science and planning. Enhanced measurement
tools and modeling concepts may open the way and the willingness to use infor-
mation technologies to address issues of social equity, imbalances in regional eco-
nomic development, and threats to the sustainability of local and global
environments.
Part I
2.1 Introduction
access within cyberspace; the possibility to substitute virtual for physical access;
the comparative quality ofthe access experience in case of substitution; and so on.
Another problem is that accessibility has often been treated as a purely spatial
issue. Yet individual scheduling of activities is not only a spatially constrained
process but one that is also strongly time dependent. More than ever before, acces-
sibility should be approached as a time-space phenomenon. While the idea is not
new (it goes back to Hägerstrand and the 1950s), it has new implications at an age
when virtually instant access to so me opportunities frees up considerable time and
thus enables access to others. The scheduling of activities may no longer be con-
strained by the spatial logic of multi-purpose trips, or the temporal constraints of
business hours. New temporal constraints come into playas we access people and
places in very different time zones across the globe - and so on. It is thus with re-
lief that we see in the paper by Forer and Huisman that the familiar time-
geography framework developed by Hägerstrand can be extended to take into ac-
count so me of these novel situations. It is unclear whether it can be equally weil
adapted to all ofthem.
Among more recently formulated concepts, a particularly useful one appears to
be that of proximal space. This defines a place as part of its (physical or func-
tional) vicinity and thus allows us to see places in the context of the other places
with which they interact (or to/from which they are accessible). In traditional geo-
graphic terminology, proximal space embodies both the site and situational char-
acteristics of locations. For example, a site may provide good bus service to a set
of other locations that together constitute its proximal space from the point of
view of physical accessibility; or it may provide the means to connect to the Inter-
net, allowing users to access specific other locations that together help form the
proximal space for these users at that particular location. How can we compare
these two instances of proximity or accessibility? In her paper (Chapter 3), Scott
makes creative use of the proximal space notion to compare accessibility in physi-
cal versus functional space in the Los Angeles area. Proximal space (characterized
by various degrees of proximity) is also implicit in Heikkila's geographic juzzy
clubs (Chapter 6), where greater accessibility (however defined) to a place is rep-
resented as a higher degree ofmembership in a corresponding 'club'.
Scott (Chapter 3) and Shen (Chapter 4) make some ofthese abstract ideas more
concrete by focusing on the role of new technologies in affecting accessibility for
employment. With still small but increasing numbers of people turning to tele-
commuting, models must be able to indicate what employment opportunities are
reachable (physically, virtually, or both ways) by a person at a given location. A
composite measure of accessibility devised by Shen takes into account the varying
mix of physical and virtual accessibility to jobs for Boston residents. The model
presented by Scott combines a traditional spatial interaction model with the Gt
local statistic devised by Getis and Ord (1992) to shed new light on the notorious
spatial mismatch phenomenon (a classic case of deficient accessibility) in the Los
Angeles area. Unlike its predecessors, Scott's model can be extended to the study
of accessibility to employment in virtual and hybrid (physical and virtual com-
bined) spaces.
18 H. Couclelis and A. Getis
As they should, the chapters in this section raise a number of questions, some
very theoretical, others very practical and technical. To what extent do we need
new concepts and measures of accessibility, rather than adaptations of existing
concepts and measures? What theoretical criteria might guide prediction or expla-
nation in accessibility research, and what technical criteria might help us choose
among different measurement approaches? Given the obvious relevance of geo-
graphie information science and technology, how can GIS functions and opera-
tions be used to represent and measure the expanded notions of accessibility
discussed here? And how can other promising formalisms (e.g., local statistics,
fuzzy set theory) enhance the usefulness of GIS for accessibility research?
A recurrent theme in the papers is the problem of scale. This is expressed as either
the micro-macro problem in a spatial context or the individual-based versus ag-
gregate measure problem more generally. Disaggregate and aggregate-level meas-
ures serve different functions; each allows for different types of questions. The
behavioral, decision-making strengths of the micro approach are often undermined
by the unavailability of individual data. The aggregate approach, while meeting
many analytical requirements (e.g., more manageable sampie sizes) is by its nature
limited to the study of problems dealing with group behavior and averages. The
age-old question of the appropriate spatial resolution continues to be debated. The
debate is complicated by the apparent space-lessness (and scale-lessness) of many
of the newer telecommunications and information technologies. Multi-scale or
scale-free approaches to analysis do not necessarily do justice to the complex ef-
fects ofthe new technologies on accessibility and human interaction.
The papers presented in this section provide some promising clues. Several of
them successfully accomplish the transition from individual to aggregate and from
local to metropolitan levels of analysis. The approach presented by Forer and Hu-
isman (Chapter 4) models individual access to educational opportunities as opera-
tional trajectories within time-space prisms, and then generalizes these into
aggregate patterns of accessibility for the student population. Heikkila's idea of
fuzzy clubs also appears to bridge the gap between individual circumstances and
the characteristics of clubs with respect to accessibility for their membership. Still,
the micro/macro problem in complex systems, as urban and regional systems in-
variably are, reaches beyond the issues of aggregation and scale. A very important
related aspect is that of emergent properties and behavior. As an example, it is
well known that the individual increase in accessibility afforded by widespread
automobile ownership in North American cities resulted in urban structures char-
acterized by low overall accessibility scores. What the corresponding emergent
phenomena for the information-age city may be, is anybody's guess. Further, by
Conceptualizing and Measuring Accessibility 19
focusing too hard on how best to aggregate individual choices and paths in space-
time, one may lose sight of the fact that these choices and paths themselves are to
a large extent socially constructed (more simply put, enabled and constrained by
aggregate behavior and its spatial consequences). Finally, we should keep in mind
that a large part of urban structure may be explained not by the accessibility needs
of individuals but by their distancing needs, as they strive to avoid the vicinity of
less desirable groups, land uses or environments. While these centrifugal tenden-
cies have always been present, the technological and organizational possibilities of
the information age may greatly amplify these phenomena.
the most important functional relations? Here Daniei Sui (Chapter 7) gives a par-
ticularly provocative answer. He argues that we may be barking up the wrong tree
by focusing on a mechanistic notion such as accessibility, when the evolutionary
notion of adaptability may be much better suited to the fluid, flexible, immaterial
and strongly cognition-oriented types of relations fostered by the information age.
To the extent that further research on accessibility is worth pursuing, there is
consensus that the focus should be on a generalized notion of accessibility as
process taking place in physical, virtual, or hybrid (physical/virtual) spaces. As
Occelli (see Chapter 17) notes, accessibility as a concept has always occupied the
intersection of physical and functional (including virtual) spaces, the place where
the socio-economic relations woven in functional space touch ground, and where
the spatial relations still constrain the functional. Studying that intersection has
always been a considerable challenge, because functional space potentially has
indefinitely more dimensions than physical (geographie) space. The difficulty in-
creases manifold in the information age with the appearance of virtual space,
which shares properties of both the physical and the functional. Thus the challenge
will be to find consistent conceptualizations that can handle the dynamic intercon-
nectedness of physical, functional, and virtual relations over space, as weB as the
infinite capacity of individuals and societies to both adapt to and modifY the ever-
changing contexts of their interactions.
References
Getis, A. and Ord, J.K. 1992. The analysis of spatial autocorrelation by use of distance
statisties. Geographical Analysis 24(3): 189-206.
3 Evaluating Intra-metropolitan Accessibility
in the Information Age: Operation al Issues,
Objectives, and Implementation
Lauren M. Scott
Environmental Systems Research Institute, 380 New York Street, Redlands CA 92373,
USA. Email: LScott@ESRI.com
3.1 Introduction
This section of the chapter reviews the accessibility literature in order to identifY
definitional and representational issues related to operationalizing an effective
analytical framework for evaluating intra-metropolitan accessibility.
Defining Accessibility
Accessibility ... is a slippery notion ... one of those COinmon terms that everyone uses
until faced with the problem of defining and measuring it (Gould 1969, 64).
A large body of formal urban theory contends that accessibility is at the very core
of processes shaping urban spatial structure: people choose residential locations
that satisfY both housing needs and workplace access, and employers choose work
locations that are accessible to employees, urban infrastructure, and consumer
markets (Giuliano 1995,3). Consequently, many researchers indicate that the con-
cept of accessibility is fundamental to definitions and explanations of urban form,
function, and efficiency, arguing that a location's access to economic and social
opportunities largely determines its value, development intensity, and economic,
social, and political uses (Wachs and Kumagai 1973,438; Knox 1980,368; Koe-
nig 1980, 169).
Intra-metropolitan Aeeessibility 23
(1) Potential vs. Outcome. At a very broad level, measures of accessibility may
be grouped into one of two definitional categories (Breheny 1978): potential
measures and outcome measures. Potential measures consider accessibility to be a
property of specific locations or individuals, and may involve counting spatial
opportunities and/or measuring distances between origins and destinations, but
they do not incorporate actual travel behavior, or use observed travel flows to
calibrate or to simulate measure components. Isochronic accessibility measures,
for example, define accessibility in terms of the total number of spatial opportuni-
ties within a specified distance or time cost of a particular location i, regardless of
whether or not individuals at i actually use these spatial opportunities. These po-
tential measures define accessibility in terms of the potential for spatial interac-
tion.
Outcome measures, on the other hand, define accessibility in terms of behavior,
as expressed through observed travel patterns. These outcome approaches consider
proof of accessibility to be a function of realized accessibility - the actual use of
services or actual participation in activities surrounding specific origins (Morris,
Dumble, and Wigan 1979, 92). Network models of accessibility relying on actual
travel flows to identifY highly accessible nodes, for example, reflect this outcome
definition of accessibility.
Spatial interaction models - the most commonly used accessibility measures -
incorporate elements of both the potential and the outcome definitional categories.
While these accessibility measures generally define accessibility in terms of the
1 See Ingram (1971), Haynes and Fotheringharn (1984), and Iones (1981).
2 For example, Hanson and Sehwab (1987) examine the link between aeeessibility and in-
dividual travel behavior using data from Uppsala, Sweden.
3 Knox (1980), for example, investigates aeeessibility to primary medieal eare in Edin-
burgh, Seotland, ultimately ereating a map showing both under- and over-servieed re-
gions.
24 L.M. Scott
potential for spatial interaction, they calibrate model parameters using actual
travel behavior (the outcome-measure strategy).
Whenever accessibility models rely on actual travel behavior, either directly or
for model calibration, it becomes troublesome to disentangle structure from
agency. Suppose, for example, we find that journey-to-work distances have in-
creased. Using an outcome definition of accessibility (i.e., using actual travel
flows), it is difficult to determine whether the longer commutes are the result of
improved accessibility (an improved transportation system, for example, may pro-
vide access to better jobs or to better hornes at a farther distance away) or whether
the longer commutes are the result of diminished accessibility (workers may be
required to travel farther because suitable employment or housing is just not avail-
able nearby) (Knox 1980, 369).
(2) Mobility vs. Accessibility. Handy (1994) notes that accessibility has only
recently become a focus in transportation planning. Traditionally, transportation
planners have emphasized mobility and infrastructure performance over concerns
about accessibility. Mobility refers to the ease of movement or the physical ability
to transcend space (facilitated travel), and encompasses monitoring the infrastruc-
ture for travel (road capacities, speed limits, and congestion, for example). Acces-
sibility, on the other hand, extends this concept of mobility, to include
examination of the context for travel (Helling 1998). Travel is rarely undertaken
for the sake of movement alone, but instead takes place within specific contexts
(Hans on 1998), motivated by the desire or need to satisfy a variety of economic,
social, or psychological objectives (Wachs and Kumagai 1973,439). The concept
of accessibility fully encompasses this notion of context, extending the scope of
concern associated with mobility to inc1ude the spatial/temporal opportunities
provided at destinations and the social, economic, political, and psychological
capability to reach destinations (Handy 1994). As a planning goal, then, a focus on
accessibility reflects a broader, more inc1usive concept that has advantages over an
exclusive focus on mobility.
4 Consider 3 isolated grid cells, each with 100 jobs. Even ifthe first cell has no workers, the
second has 100 workers, and the last has 1000 workers, scores for all three cells will be
equal, for many accessibility models.
Intra-metropolitan Accessibility 25
Modeling Accessibility
5 Where site characteristics reflect the attributes and qualities of particular locations or spa-
tial objects, situation characteristics reflect each location's embeddedness within a
broader spatial structure involving other locations and other spatial objects.
26 L.M. Scott
tive space6 . Couc1elis (1991; 1997) has developed the idea of proximal space as a
way to bridge these two concepts. Where absolute space emphasizes the locational
coordinates and attribute characteristics of site, and relative space emphasizes the
spatial relations associated with situation, the key notion in proximal space is the
neighborhood (Couc1elis 1997, 170). This notion of neighborhood - neither static
nor restricted to physical contiguity - reflects Sheppard's extended concept of
local context.
Proximal space emphasizes spatial dependenee in the form of loeal interactions
where loeal interactions may be relevant from three different perspeetives: spatial
proximity, funetional proximity, and statistical spatial dependency. Spatial prox-
imity reflects the physieal spatial relationships associated with a loeal site and its
neighbors. Funetional proximity reflects relationships based on influence or inter-
action. Statistieal spatial dependence refleets the eohesion and homogeneity of a
site and its neighbors.
Most spatial interaction aeeessibility models incorporate both a site and a situa-
tion eomponent (typically a spatial opportunities variable and a distance or imped-
anee variable). They are inadequate for representing intra-metropolitan
aecessibility and the proximal space construet, however, on at least two grounds:
(1) Sc ale. As with GIS operations, most spatial interaction models present intra-
metropolitan aceessibility as a single score or indicator for each loeation of inter-
est. This score reflects a loeation's absolute site attributes on the one hand, and its
situation relative to other loeations and their attributes on the other. The eoneept of
accessibility, however, is more realistically represented as a proeess or function of
spaee, time, and technology. Individuals may trade-off time and distanee, or may
utilize telecommunieations technologies, to gain aecess to spatial opportunities at
a distance. Space, time and available technology, therefore, funetion as struetures,
which both constrain and enable human aetivities. With spatial interaction aeees-
sibility models, these struetures are implemented by discounting spatial opportuni-
ties by an impedanee faetor. For any given location i, the result is a single
indieator representing so me quantity of whole opportunities assoeiated with loea-
tion i itself, plus some quantity of partial opportunities distributed various dis-
tanees away. Interpretation is awkward. This strategy of diseounting spatial
opportunities is not entirely effective for representing the spatial dynamies of in-
tra-metropolitan aecessibility with ehanges in seale or with movements through
spaee.
6 From the absolute or Newtonian perspective, space is represented as a distinct entity with
characteristics similar to a system of pigeonholes or containers (Lawton 1983, 197). It is
conceptualized as emptiness - an entity with existence independent of matter, possessing
the structure to hold or to individuate phenomena - a universal receptacle in which objects
exist and events occur (Smith 1984; Harvey 1973). From the relative or Liebnitz perspec-
tive, on the other hand, space is an abstract concept reflecting the spatial relationships be-
tween perceived objects, endowed with structure and properties that are intimately tied to
process (Lawton 1983, 197; Couclelis 1992,221).
Intra-metropolitan Accessibility 27
Context
7 Couclelis (1994) coined the term spatial technologies in reference to the bundle oftrans-
portation, communication, and information technologies that, in conjunction, modify spa-
tial relations.
28 L.M. Scott
ing urban spatial structure 8 and in how these impacts affect intra-metropolitan
accessibility. The proposed analytical framework, therefore, presents an aggre-
gate-level analysis rather than focusing on impacts associated with any one indi-
vidual (disaggregate-Ievel analysis). These broad structural changes have
important implications for urban and transportation planning. In addition, results
from this type of aggregate-level analysis can provide an effective springboard for
more detailed, disaggregate-Ievel analyses of intra-metropolitan accessibility.
This section of the chapter presents an analytical framework for measuring intra-
metropolitan accessibility. The framework is based on a level-of-service definition
of accessibility, the Couclelis (1997) proximal space construct, and the Getis/Ord
Gt statistic (Getis and Ord 1992, Ord and Getis 1995).
Conceptualizing Accessibility
8 In theory, urban spatial structure may be defined in terms ofthe spatial relationships link-
ing a region's urban activities (employment, schools, medical facilities, public services,
and/or recreational activities) to the spatial distribution of a region's inhabitants (Pred
1977, 10; Simpson 1987, 120). In the examples presented in the next section, however, I
limit this definition of urban spatial structure to a focus on the spatial relationships be-
tween employment opportunities and workers by place of residence. I emphasize accessi-
bility to employment opportunities because employment represents a fundamental
component of the urban landscape and has been directly impacted by the broad spatial
processes of economic restructuring, suburbanization, and rapid technological deve1op-
ments.
Intra-metropolitan Accessibility 29
to space, time, and available teehnology. The empirieal analyses presented in the
next seetion, therefore, define loeal interactions - the proximal spaee eonstruet -
using a variety of travel eosts. This approach is adopted in order to highlight the
spatial dynamies of intra-metropolitan aeeessibility assoeiated with ehanges in
seale, and to refleet the idea that individuals may trade-off time and distanee to
gain aeeess to opportunities at a distanee. The statistieal framework used to meas-
ure both struetural and funetional aeeessibility, and to model the proximal spaee
eonstruet is deseribed next.
The Getis/Ord Gi' statistie measures the degree of association or spatial clustering
assoeiated with a single variable (X) distributed over a spatial surfaee (Getis and
Ord 1992; Ord and Getis 1995). Consider an urban landscape divided into n re-
gions, i = 1, 2, ... , n, where eaeh region is identified with a loeation whose Carte-
sian coordinates are known. Eaeh loeation i has assoeiated with it a value Xi of the
variable X. Eaeh value, Xi, is affiliated with a set of neighbors - a set of Xj values -
where the neighborhood is defined by a spatial weights matrix (typieally a binary,
one/zero matrix). The spatial weights matrix is square, with row and eolumn en-
tries for eaeh pair of loeations i and j. As one example, a binary spatial weights
matrix may define loeal interactions - the neighborhood - in terms of physieal
distanee: a 2-mile-distanee radius, for example. Eaeh row, then, would be associ-
ated with eolumn entries of either one or zero: values of one would refleet mem-
bership in the neighborhood, and would be assoeiated with all loeations j within a
2-mile-distanee radius of loeation i, including the diagonal where j=i (footnote9 ).
The Gi' statistie may be written as follows (Ord and Getis 1995):
where:
Gi'(d) = the Gi' score for a partieular distanee (or other neighborhood member-
ship eriteria), d.
wij{ d) = the spatial weights matrix for a partieular distanee (or other neighbor-
hood membership eriteria), d
Xj = the spatial variable, X, being measured at loeation j, where j may equal i
l--4J' = the sum of eolumn entries for row i ofthe spatial weights matrix
X = the mean for all Xj observations
9 The Gi" and Gi formulations differ in the way the neighborhood is defined. With Gi' j may
equal i; for the Gi formulation, however, the spatial weights matrix has a zero for each di-
agonal entry. This distinction results in slight differences in the way the Gi" and Gi formu-
lations are ca1culated (see Ord and Getis 1995).
30 L.M. Scott
The Gi< statistic compares local neighborhood averages to the global average for
a given variable X in a defined study region. Where high values of X (high in rela-
tion to the global me an) cluster together, Gi< scores will be positive; where low
values of X cluster together, Gi< scores will be negative. These Gi< scores may be
interpreted as standard normal deviations where the expectation is zero and the
variance is one (Ord and Getis 1995). To understand why the Gi< statistic is effec-
tive for measuring intra-metropolitan accessibility, however, it is necessary to
consider how perfeet accessibility might express itself on the urban landscape, and
further, to consider how specific aspects of the urban landscape may serve to de-
tract from this notion of perfection.
Whenever we deal with accessibility, we are implicitly confronting a situation
involving both supply and demand. When measuring accessibility to employment
opportunities, however, this distinction becomes blurred. From the employers'
perspective, demand is reflected by a need for workers to fill particular jobs, and
the number ofworkers available represents supply. From the workers' perspective,
demand reflects a need for jobs, while the number of jobs available represents
supply. One approach to resolution is to measure accessibility by level-of-service.
Suppose we define a spatial opportunities variable (X) to reflect both jobs and
workers ll .
Let:
Notice that if location i contains 10 percent of the region's jobs and 10 percent
of the region's workers, Xi will equal o. If, however, location i contains a larger
proportion of jobs than workers, Xi will be positive. Positive Xi values reflect loca-
10 The spatial weights matrix need not be binary. When it is binary, S1i< is equal to vv;<.
11 This 'level-of-service' definition for accessibility is appropriate because the relationship
between jobs and workers is represented as a simple one-to-one correspondence. Measur-
ing accessibility to other types of services (retail or medical facilities, for example), may
require different level-of-service formulations.
Intra-metropolitan Accessibility 31
tions where employers have less than perfect accessibility to potential employees.
Similarly, negative Xi values reflect locations offering workers less than perfect
accessibility to potential employment opportunities. By defining accessibility in
terms of how weil a particular location serves surrounding populations, both the
employers' and the workers' perspectives on accessibility are represented l2 .
This notion of perfeet accessibility needs further clarification, however, as it
provides a primary justification for using the Gi' statistic - a measure of spatial
association - to measure intra-metropolitan accessibility. Imagine a highly urban-
ized hypothetical landscape exhibiting the following, quite unrealistic, characteris-
tics: (1) all jobs and all workers are evenly distributed throughout the study
region; (2) each location i is connected to every other location j by an equivalent
and effective transportation network; and (3) all jobs and workers within the study
region have identical characteristics: workers have identical preferences, levels of
mobility, skills, etc., and jobs offer equivalent working environments and wages.
In this hypothetical landscape, no matter what our scale of analysis, or how space
is partitioned into discrete zones (census tracts, for example), worker proportions
will match job proportions and the spatial opportunities variable (X), as defined
above, will consistently tend toward zero. This hypothetical landscape reflects a
notion of 'perfect' accessibility defined in terms of level-ofservice or social eq-
uity; each worker in the study area has similar accessibility to the region's em-
ployment opportunities. Now let us relax each of the assumptions given above.
Urban landscapes, generally, are not associated with evenly distributed jobs and
workers. Instead, they are typically structured into a mosaic of residential and in-
dustrial land uses. Where job-rich (worker-poor) sites intermingle with worker-
rich (job-poor) sites, worker shortages at one location may balance job shortages
at locations nearby, so that accessibility is maintained. Where job-rich or worker-
rich sites cluster together, however, worker/job shortages are additive. Scale of
analysis is important here. In addition, transportation and communications net-
works develop unevenly over time and space. Where residential neighborhoods
and employment centers are efficiently connected via transportation or telecom-
munications networks, costs 13 to access opportunities at a distance will be smalI;
where there are physical or functional barriers restricting movement or communi-
cations, opportunity costs will be large. Finally, employment outcomes (a firm's
decision to hire an individual or a worker's decision to accept a job offer) are in-
fluenced by a large number of complex individual characteristics and behavioral
factors associated with worker or firm preferences, motivations, aspirations, objec-
tives, social networks, search strategies, and sometimes just plain dumb luck. One
12 This level-of-service variable does not include information regarding commuting pat-
terns.
13 There are, of course, costs associated with operating a vehicle or using public transporta-
tion to overcome distance. Other, traditional costs, may involve the time needed to tran-
scend space (walking to work). Even when telecommunications technologies are
employed, however, there are costs associated with gaining access to appropriate equip-
ment, education, and authority to use these technologies. Many employees, who choose
to telecommute to the workplace, may pay for this convenience in the form of reduced
opportunities for promotion or salary increases as a result of spending large amounts of
time off-site.
32 L.M. Scott
approach to understanding how these complex individual factors play out in the
aggregate, however, is to examine observed journey-to-work travel behavior.
The Gi' statistic presents an effective approach for evaluating intra-metropolitan
accessibility by considering each location i within the context of its proximal loca-
tions j, and determining the degree of spatial clustering associated with job and
worker spatial patterns at multiple scales of analysis. The proximal space relations
reflecting the complex physical and functional connections among sites in the
study area are modeled via the Gi' statistic spatial weights matrix. In the empirical
analyses presented in the next section, the spatial weights matrix is modeled to
reflect Euclidean distances, travel time costs, and functional time costs at multiple
scales of analysis. Functional travel times, derived from actual journey-to-work
travel patterns, are compared to actual travel times, providing insight regarding the
impact changes in transportation and telecommunications technologies are having
on urban spatial structure. These statements are best clarified through the exam-
pies presented in the next section.
3.4 Application
Greater Los Angeles (Figure 3.1) is generally viewed as a region of endless urban
sprawl with widely dispersed population and employment patterns (Giuliano and
Small 1991); the region has been described as the prototypical example of urban
restructuring (Soja, Morales, and Wolff 1983). This study area, therefore, provides
an especially appropriate context for investigating questions relating to intra-
metropolitan accessibility. In this final section of the chapter, the proposed ana-
lytical framework is applied to 1990 employment data for the five county Greater
Los Angeles region 14. Three analyses are presented. While each is performed us-
ing the framework outlined above, the method used to model the relations con-
necting origins (workers by place ofresidence) and destinations Gobs) varies.
14 The study area encompasses most of the urbanized portions of the five-county Greater
Los Angeles region covered by the Southem Califomia Association of Govemments
(SCAG): Los Angeles, Orange, San Bemardino, Riverside and Ventura counties. Em-
ployment data reflecting the number of jobs and the number of resident workers for each
census tract in 1990 were kindly provided by SCAG. Travel-time data and joumey-to-
work travel-flow data were obtained from the 1990 Census Transportation Planning
Package (CTPP) CD-ROM distributed by the V.S. Bureau ofTransportation Statistics.
Intra-metropolitan Accessibility 33
~. j---- ,.
Oxna~~\
Spatial Distributions
for local workers. Tracts with scores near zero represent regions offering effective
accessibility to both employers and workers l5 .
:~ !:iI~i~!Et.
... ··::::I:i:lr~::::·:··:::?\:· . : : :.:
•
Accessibility:
III! Very Poor for Workers
::::::::::::::::::::: Poor for Workers
Effective
Poor for Employers
. . Very Poor for Employers
Figure 3.2. Greater Los Angeles, 1990. Accessibility scores based on the spatial distribu-
tion of jobs and workers. Scale of analysis: 5 miles.
15 In Figures 3.2, 3.3, 3.4, and 3.6, accessibility designations of very poor are associated
with Gi" scores greater than +2, or less than -2, standard deviations; designations of poor
are associated with Gi" scores ranging from + 1 to +2 or from -1 to -2. Effective ace es si-
bility is associated with Gi" scores between -1 and + 1.
Intra-metropolitan Accessibility 35
Accessibility:
~ Very Paar far Warkers
i:~'It~/);/
::JJt11 Paar far Warkers
Effective
Paar far Em players
. . Very Paar far Emplayers
Figure 3.3. Greater Los Angeles, 1990. Accessibility scores based on the spatial distribu-
tion of jobs and workers. Scale of analysis: 10 miles.
----CD
.'.~.'.
:.:"-:.::;::~l:::
Accessibility:
~ Very Paar for Warkers
ttt:U}: Paar far Warkers
Effective
Paar far Emplayers
Very Paar far Emplayers
Figure 3.4. Greater Los Angeles, 1990. Accessibility scores based on the spatial distribu-
tion of jobs and workers. Scale of analysis: 15 miles.
36 L.M. Scott
if workers are willing and able to endure 15-mile commutes. At the same time,
however, worker-rich Gob-poor) clustering in the eastern and northern portions of
the study region have become more intense: the job deficiencies in these regions
are additive with increases in scale of analysis.
Analysis is performed at 5, 10, and 15 miles to emphasize the idea that accessi-
bility is not a static score, but a function, or process of space, time, and technolgy.
Consider, for example, the two tracts labeled 1 and 2 in Figure 3.4. Both tracts are
associated with similar accessibility scores when evaluated for a 15-mile distance
radius. Considering only this one score, however, obscures important details about
variations in accessibility with changes in scale. Figure 3.5 graphs the Gi' scores at
5, 10, and 15 miles. For the Orange County tract (labeled 2), accessibility to em-
ployment opportunities increases rapidly if one is willing and able to travel 10 or
15 miles. For the Los Angeles County tract (labeled 1), however, increased travel
offers diminishing returns. Getis (1994) suggests one approach to capturing these
scale-related variations in Gi' scores is to report not only the Gi' scores at multiple
spatial seal es, but to also report the slope associated with these scores.
2.0
1.0
r_ _ _
I/)
~
8
CI) 0.0
iC
c:5 2_______ .. - - - - - --
-1.0
--- • LA Tract
- -. - -oe Tract
-2.0
5 10 15
Miles
Figure 3.5. Variations in accessibility with changes in scale.
Transportation Infrastructure
Accessibility:
Very Poor for Warkers
:IItII Poar for Warkers
Effective
Paar for Employers
Very Poor far Emplayers ....:..
Figure 3.6. Greater Los Angeles, 1990. Accessibility scores based on travel-time costs.
Scale of analysis: 30 minutes
16 Median tract-to-tract travel times were extracted from the CTPP data set. Unfortunately,
these data are of poor quality. Not only did a large number of missing time costs need to
be estimated, but approximately 10 percent of the time values had to be discarded alto-
gether because they were clearly erroneous (indications that individuals driving alone in a
car could travel approximately 90 miles in 15 minutes, or required 99 minutes to travel
approximately 0.5 miles). Where necessary, travel time estimates were constructed from
similar, plausible commute times (similar distances and proximal origins and destina-
tions). The large number of estimates needed, however, limits the extent to which analy-
sis is possible. These estimates are sufficient, however, for demonstrating the potential,
given accurate data, to address a broad range ofurban research questions using the meth-
ods outlined in this chapter.
38 L.M. Scott
Transportation:
IIIlnhibits Accessibility
Has little Impact
~ Facilitates Accessibility
Figure 3.7. Greater Los Angeles, 1990. Transportation network impacts on accessibility.
Scale of analysis: 30 minutes.
Striking differences are found in the northern portion of the study area, where
accessibility based on distance indicates poor accessibility for workers, but based
on travel times indicates poor accessibility for employers. In this region, effective
transportation networks allow workers in the San Fernando Valley access to job-
rich tracts to the south. Figure 3.7, based on differences in accessibility scores
using both the 15-mile-distance radius and the 30-minute-time-cost isochrone,
indicates how the transportation network impacts intra-metropolitan accessibility
at this scale of analysis. Tracts with a block fill pattern indicate regions with con-
gestion or insufficient transport capacity, reducing accessibility potential. Tracts
with a striped fill pattern indicate regions where transportation infrastructure is
facilitating accessibility. The accessibility scores associated with tracts shown
using a smooth fill pattern did not change as a result of factoring in transportation
infrastructure infl uences.
Analyses using travel-time costs allow evaluation and isolation of the impacts
on intra-metropolitan accessibility imposed by transportation infrastructure, and
offer some very interesting possibilities for extended analysis. Figure 3.7 indicates
where insufficient infrastructure is limiting accessibility potential. An alternative
to using travel-time costs is using actual road networks and their associated capac-
ity constraints in constructing the spatial weights matrix. With this type of model,
it would be very interesting to test whether the congestion depicted in Figure 3.7
would be most effectively relieved by building more roads or, alternatively, by
Intra-metropolitan Accessibility 39
Functional Proximity
A final example of the Gi' analytical framework represents the relations between
origins and destinations as functional time costs. Functional time costs are one
instance of what Gatrell (1983) refers to as interaction proximities. The concept of
functional time costs or functional distances is based on the idea that two places
with high rates of spatial interaction are functionally 'closer' than two places with
very little interaction. Participation in place-based activities promotes familiarity,
which may be expressed through extended social networks, development of stra-
tegie contacts, and/or expanded knowledge about a place or region (see Hanson,
1998). In a recursive manner, participation in place-based activities promotes fa-
miliarity, while familiarity increases the likelihood of participation in place-based
activities. Using a similar logic, two places with very limited spatial interaction
may be represented as being functionally distant.
To operationalize the concept of functional travel times, I adapt a method devel-
oped by Tobler and Wineburg (1971) in an article titled A Cappadocian Specula-
tion. The adapted method involves inverting the doubly constrained gravity model
with known journey-to-work flows to solve for travel time. The doubly con-
strained gravity model may be written as:
Inverting the model and using known values for Fij, we may solve for tij to obtain
functional time:
1fb
Functional Accessibility:
_ Dominated by Interaction
_ with Worker-Rich Tracts
Effective
~ Dominated by Interaction
~ with Job-Rich Tracts
Figure 3.8. Greater Los Angeles, 1990. Accessibility scores based on functional travel
times. Scale of analysis: 30 minutes.
Intra-metropolitan Accessibility 41
Figure 3.9. Greater Los Angeles, 1990. Comparison of actual travel times to functional
travel times.
Where actual spatial interaction rates are higher than predicted by the gravity
model, functional travel times will be shorter than actual travel times; where spa-
tial interaction rates are lower than predicted, functional times will be longer than
actual travel times. Figure 3.8 maps the Gi' accessibility scores calculated using a
spatial weights matrix based on a 30-minute functianal time-cost isochrone. This
map is quite different from the others. Notice that many of the tracts near Down-
town Los Angeles, while physically c10ser to job-rich (worker-poor) tracts, are
functionally c10ser to worker-rich tracts. Similarly, the job-poor tracts in south
Orange County are functionally c10ser to job-rich tracts to the north than to the
worker-rich tracts physically nearby. Workers in south Orange County are, there-
fore, exposed to a large number of jobs because of their commuting behavior. No-
tice, also, that Figure 3.8 is not a camplele reversal ofFigure 3.6. The worker-rich
tracts in the eastern-most portion of the study area, for example, continue to be
associated with worker-rich clusters. One interesting extension of this analysis
would be to actually map the stretching and pulling of space depicted by the func-
tional travel times used in the Figure 3.8 analysis. An alternative analysis is pre-
sented in Figure 3.9. Here the average functional travel times for all journey-to-
work flows are compared to the average actual travel times for these same flows.
Where functional travel-time averages are similar to actual travel-time averages,
we may conclude that accessibility is primarily determined by structural compo-
nents of accessibility: the spatial distribution of jobs in relation to resident work-
ers, and the impacts imposed by transportation infrastructure. However, tracts
associated with much shorter functional travel-time averages, indicate that non-
structural influences are impacting accessibility. These non-structural influences
might include higher incomes, effective access to private vehicles, association
42 L.M. Scott
Accessibility:
. . Poor for Workers
Effective
Poor for Employers
Inconsistent
Figure 3.10. Greater Los Angeles, 1990. Accessibility score stability. Based on: 15-mile
distance, 3D-minute travel time, and 3D-minute functional time analysis.
Intra-metropolitan Accessibility 43
3.5 Conclusions
This chapter has reviewed some of the definitional and operational challenges that
must be met if we are to effectively examine impacts on intra-metropolitan acces-
sibility associated with the broad spatial processes of economic restructuring, sub-
urbanization, and rapid developments in transportation and telecommunications
technologies. It has outlined an analytical framework structured around a level-of-
service definition of accessibility, the Couclelis proximal space construct, and the
Getis/Ord Gi" statistic. Future research will consider use of actual road networks,
incorporation of highway capacity constraints to develop a dynamic accessibility
model, exploration of the efficacy of defining the relations between origins and
destinations using other attribute and interaction proximities, and consideration of
other methods for deriving functional travel times. Taking analysis of intra-
metropolitan accessibility in the Greater Los Angeles region to the next level, will
require accurate data over aseries oftime periods. While the challenges are daunt-
ing, understanding how broad spatial processes in the information age shape urban
spatial structure represents one of the most interesting and exciting tasks facing
urban geographers at this time.
References
Arentze, T.A., Borgers, A.WJ., and Timmermans, HJ.P. 1994. Multistop-based measure-
ments of accessibility in a GIS environment. International Journal of Geographical In-
formation Systems 8:343-56.
Breheny, MJ. 1978. The measurement of spatial opportunity in strategie planning. Re-
gional Studies 12:463-79.
Bums, L.D. 1979. Transportation, temporal, and spatial components of accessibility. Lex-
ington: Lexington Books.
Chinitz, B. 1991. A framework for speculating about future urban growth patterns in the
USo Urban Studies 28:939-59.
Couclelis, H. 1991. Requirements for planning-relevant GIS: a spatial perspective. Papers
in Regional Science 70:9-19.
Couclelis, H. 1992. Location, Place, Region, and Space. In Abler, R.F., Marcus, M.G., and
Olson, J.M. (eds.) Geography's Inner Worlds: Pervasive Themes in Contemporary
American Geography. New Brunswick, NI: Rutgers University Press, 215-33.
Couclelis, H. 1994. Editorial: Spatial Technologies. Environment and Planning B: Plan-
ning and Design 21: 142-43.
Couclelis, H. 1997. From cellular automata to urban models: new principles for model de-
velopment and implementation. Environment and Planning B: Planning and Design
24:165-74.
Gatrell, A.C. 1983. Distance and Space: a Geographical Perspective. Oxford: Clarendon
Press.
Geertman, S.C.M. and Ritsema Van Eck, lR. 1995. GIS and models ofaccessibility poten-
tial: an application in planning. International Journal of Geographical Information Sys-
tems 9:67-80.
44 L.M. Seott
Getis, A. 1994. Spatial Dependence and Heterogeneity and Proximal Databases. In Fother-
ingharn, S. and Rogerson, R. (eds.) Spatial Analysis and GIS. London: Taylor & Francis,
213-29.
Getis, A. and Ord, J.K. 1992. The analysis of spatial assoeiation by use of distanee statis-
ties. Geographical Analysis 24: 189-206.
Giuliano, G. 1995. The weakening transportation-Iand use eonneetion. Access, 6:3-11.
Giuliano, G. and SmalI, K.A 1991. Subcenters in the Los Angeles Region. Regional Sci-
ence and Urban Economics 21: 163-82.
Gould, P. 1969. Spatial Diffusion: Commission on College Geography. Washington DC:
Association of Ameriean Geographers.
Graham, S. and Marvin, S. 1996. Telecommunications and the City: Electronic Spaces,
Urban Places. London: Routledge.
Handy, S.L. 1994. Highway blues: nothing a little aecessibility cantt eure. Access 5:3-7.
Hanson, S. 1998. Off the road? Refleetions on transportation geography in the information
age. Journal ofTransport Geography 6:241-49.
Hanson, S. and Schwab, M. 1987. Accessibility and intraurban trave\. Environment and
Planning A 19: 735-48.
Harvey, D. 1973. Social Justice and the City. Baltimore: The Johns Hopkins University
Press.
Haynes, K.E. and Fotheringharn, AS. 1984. Gravity and Spatial Interaction Models. Bev-
erly HiIIs: Sage Publications.
Helling, A. 1996. Why We Should Care About Intra-Metropolitan Accessibility and How
We Measure It. In Couclelis, H. (ed.) Spatial Technologies, Geographie Information, and
the City Proceedings (Baltimore, Maryland, Sep 9-11). Santa Barbara CA: National Cen-
ter for Geographie Information and Analysis.
Helling, A. 1998. Changing intra-metropolitan aeeessibility in the U.S.: evidence from
Atlanta. Progress in Planning 49:55-107.
Ingram, D.R. 1971. The eoneept of aeeessibility: a seareh for an operational form. Regional
Studies 5:101-7.
Jones, S.R. 1981. Accessibility Measures: a Literature Review. Transport and Road Re-
search Laboratory Report 967. Crowthome, Berkshire: TRRL.
Knox, P.L. 1980. Measures of aecessibility as social indieators: a note. SocialIndicators
Research 7:367-77.
Koenig, J.G. 1980. Indicators ofurban aecessibility: theory and application. Transportation
9: 145-72.
Kwan, M. 1998. Spaee-Time and integral measures of individual aceessibility: a compara-
tive analysis using a point-based framework. Geographical Analysis 30: 191-216.
Lawton, R. 1983. Spaee, place and time. Geography 68:193-207.
MiIler, HJ. 1991. Modelling aeeessibility using spaee-time prism eoneepts within geo-
graphieal information systems. International Journal of Geographical Information Sys-
tems 5: 287-301.
Morris, J.M., Dumble, P.L., and Wigan, M.R. 1979. Aeeessibility indieators for transport
planning. Transportation Research A BA: 91-109.
Ord, J.K. and Getis, A 1995. Local spatial autoeorrelation statisties: distributional issues
and an applieation. Geographical Analysis 27: 287-306.
Pirie, G.H. 1979. Measuring aeeessibility: a review and proposa\. Environment and Plan-
ning A 11 :299-312.
Pred, A. 1977. City-Systems in Advanced Economies. London: Hutehinson.
Intra-metropolitan Accessibility 45
Sheppard, E. 1996. Commentary: Site, situation, and social theory. Environment and Plan-
ning A 28: 1339-44.
Simpson, W. 1987. Workplace location, residential location, and urban commuting. Urban
Studies 24: 119-28.
Smith, N. 1984. Uneven Development: Nature, Capital, and the Production ojSpace. Ox-
ford: Basil Blackwell.
Soja, E., Morales, R., and Wolff, G. 1983. Urban restructuring: an analysis of social and
spatial change in Los Angeles. Economic Geography 59:195-230.
Tobler, W.R. and Wineburg, S. 1971. A Cappadocian speculation. Nature 231 :39-41.
Wachs, M. and Kumagai, T.G. 1973. Physical accessibility as a social indicator. Socio-
Economic Planning Sciences 7:437-56
4 Transportation, Telecommunications, and
the Changing Geography of Opportunityl
Qing Shen
4.1 Introduction
I Reprinted with permission from Urban Geography, Vol. 20, No. 4, pp. 334-355. ©V.H.
Winston & Son, Ine., South Oeean Boulevard, Palm Beaeh, FL 33480. All rights reserved.
48 Q. Shen
an urgent need for new concepts and methods that will allow urban researchers to
examine the new geography as an interdependent whole (Couc1elis 1996).
This chapter aims to achieve two objectives. The first is to develop an analytical
framework that integrates location, transportation, and telecommunication vari-
ables into a unified representation of spatial relationships. The second objective is
to use the new analytical framework as a tool to undertake structured explorations
of spatial and social implications oftelecommunications. The focus ofthis study is
on the changing geography of opportunity in American metropolitan areas.
The next section presents the new analytical framework, which is developed
based on literature in three areas of urban research. The first is spatial representa-
tion using accessibility measures (Morris, Dumble, and Wigan 1979, Shen 1998a;
Weibull 1976), the second is urban industrial restructuring (Kasarda 1995, Office
of Technology Assessment, U.S. Congress 1995), and the third is telecommuting
(Mokhtarian 1990, Salomon and Mokhtarian 1997). The third section will apply
the new analytical framework to an analysis of the changing geography of oppor-
tunity. The discussion starts with observations of spatial reconfigurations that have
taken place in metropolitan areas that have been shaped by the automobile. It is
then expanded by taking into consideration new options of access, new composi-
tions of economic opportunities, and new spatial patterns of metropolitan growth
that are emerging in the information age. The fourth section will discuss some
findings of a case study of employment accessibility in the Boston Metropolitan
Area. The analysis of existing patterns and the simulation of future change both
shed light on the understanding of the spatial and social effects of new telecom-
munications. In the conciuding section, some policy implications will be drawn
and directions for future research will be identified.
Basic Concepts
The discussion in this chapter involves several important concepts. One such con-
cept is spatial technologies, which is defined by Coucielis (1996) as the complex
of transportation, communication, and information technologies that together
modifY spatial relations. Another concept is accessibility, which is a measure of
the strength and extensiveness of spatial relationships between opportunity seekers
and relevant opportunities. Accessibility is a basic concept in geography, because
it indicates the potential for spatial interaction. A third concept is geography 01
opportunity, which can be defined as a collection of people, economic opportuni-
ties, and the spatial relationships between them. In many studies, inciuding this
one, the metropolitan area is considered as the territorial extent of the geography
of opportunity.
The Changing Geography of Opportunity 49
The general relationships among these three concepts are clear. Spatial tech-
nologies modifY spatial relations, and the effects can be measured concretely by
examining changes in accessibility. Changes in accessibility, in turn, imply possi-
ble reconfigurations of the geography of opportunity. These relationships become
specific and meaningful when examined in some real context, such as a metropoli-
tan area. They become much more complex once the social dimension is added to
the analysis.
Over the decades, urban researchers have developed accessibility measures to help
und erstand the complex relationship between transportation technologies and the
changing geography of opportunity. Their efforts have resulted in a large volume
of literature. Most researchers agree that the level of accessibility for an opportu-
nity seeker is a function of (4.1) the total number of relevant opportunities, (4.2)
the spatial distribution of these opportunities, (4.3) the spatiallocation of the indi-
vidual, and (4.4) the individual's ability to overcome spatial separation. When
opportunities are relatively scarce, the competition with other opportunity seekers
must also be taken into account. Many mathematical formulations have been pro-
posed for the measurement of accessibility (Morris, Dumble, and Wigan 1979). In
the following formulation, first suggested by Weibull (1976) and recently general-
ized by Shen (l998a), each available opportunity is weighted by a demand factor,
expressed as the denominator, which represents the competing demand generated
by other opportunity seekers:
A; = ~ - - - - - - - (4.1 )
where:
A; is the accessibility for opportunity seekers who live in zone i and
travel by mode v
Oj is the number of relevant opportunities in zone j
f(tgV) is the impedance for travel from i to j by mode v
f(tkjm) is the impedance for travel from k to j by mode m
p km is the number of opportunity seekers who live in zone k and travel by
modem
For a metropolitan area with N zones, i, j, k = 1, 2, ... , N
For a metropolitan area with M modes, v, m = 1, 2, ... , M
50 Q. Shen
(4.2)
where:
A~ is the general aeeessibility for all opportunity seekers who live in
zone i
P; is the number of opportunity seekers who live in zone i and travel by
mode v
Pi is the total number of opportunity seekers in zone i
Opportunity seekers can be divided into two broad categories based on whether
or not they have telecommunications capabilities:
2 It is important to note that the classification presented here is different from what was
described in an earlier paper by this author (Shen 1998b). The old classification included
only two categories of opportunities - opportunities that can only be accessed through
telecommunications and opportunities that can only be accessed through transportation.
The new classification has an additional category, which consists of opportunities that can
be accessed through either transportation or telecommunication. The new scheme is more
sophisticated and realistic. Because of the difference in classification, the resulting
formulations and measurements of accessibility are also different in some ways.
The Changing Geography ofOpportunity 53
(4.3)
(4.4)
where:
At(1) is accessibility to opportunities that are accessed through transporta-
tion only;
A;V(2) is accessibility to opportunities that can be accessed through either
transportation or telecommunications, as measured for people who
can use transportation only;
Aj cv(2) is accessibility to opportunities that can be accessed through either
transportation or telecommunications, as measured for people who
can use telecommunications;
54 Q. Shen
(4.5)
In other words, there are two variables that determine the impedance for an op-
portunity seeker who has telecommunications capabilities. One is the frequency of
using transportation for opportunity-seeking trips, and the other is the zone-to-
zone travel times for the transportation mode used. Note that (]' is a parameter that
converts an actual average daily travel time into a perceived average daily travel
time. 3 There are M different specifications of impedance for people who have tele-
communications capabilities, each corresponding to a transportation mode. To
keep the discussion simple, it is assumed that the travel frequency for people who
use telecommunications to aeeess opportunity does not vary aeross geographie
loeations or types of opportunities.
There are many possible mathematical formulations for specifying equations 4.3
and 4.4. Here the discussion is based on the assumption that equation 4.1, de-
scribed previously, is a proper representation of accessibility in the physical space,
and can be extended to represent more eompletely accessibility in the emerging
geography ofthe information age.
Accessibility for different groups of opportunity seekers can be measured using
the following equations, which are the extended versions of equation 4.1:
3 The value of (J" should normally be determined through calibration. But because the em-
pirical data required for calibration are not currently available, (J" is given the value of 1
by assuming that the perceived average daily travel time equals the actual average daily
travel time.
The Changing Geography ofOpportunity 55
A; = ~ - - - - - - - + ~ - - - - - - - - - - - - - - (4.6)
ImIk p km f(t/<jm) ImIk [(1-t5JPkm f(t/<jm) + t5?t f(t/<jctn)]
A;-' = ~ - - - - - - - + ~ - - - - - - - - - - - - - -
ImIk p km f(t/<jm) ImIk [(1-t5JPkm f(t/<jm) + t5?t f(t/<jctn)]
+~------- (4.7)
Although these two equations appear quite complicated, they are in fact rather
straightforward to understand. The accessibility for opportunity seekers who do
not have telecommunications capabilities, which is represented by equation 4.6,
has two components. The first component is essentially the same as the right side
of equation 4.1, except that only (1-~-)")Oj opportunities in each zone are now
accessed exc1usively through transportation. These opportunities are accessible to
all opportunity seekers (ImIkPkm). Hence the denominator, which represents the
demand factor, includes competing demands generated by all of them. The second
component calculates accessibility attributed to ).,/0opportunities in each zone that
are accessible both to people with telecommunications capabilities (ImI kt5?km)
and people with only transportation (ImIJ1-t5JP km). The demand factor of this
component, which is the denominator, consists of competing demands generated
by these two broad categories of opportunity seekers. It is important to note that
the two impedance functions contain different travel times: travel time is t/<jctn is for
people with telecommunications; it is t/<jm is for people with only transportation.
The accessibility for opportunity seekers who have telecommunications capa-
bilities, which is represented by equation 4.7, has three components. The first
component is identical to the first component in equation (4.6), because they can
also use transportation to access the (1-~-)") Oj opportunities in each zone. The
second component is also essentially the same as the second component in equa-
tion 4.6, except that tr t/
instead of is used to specif)r the impedance function in
the numerator. The third component calculates accessibility attributed to the rjJ.o.
} }
56 Q. Shen
(4.8)
Notice that although all opportunity seekers in zone i can use transportation as
the only means of access, only LiP;-Pr;'J of them actually have to do so. This
weighted measure, similar to equation (4.2), shows that the higher the percentage
of a zone's residents that use advanced spatial technologies, the higher the zone's
general accessibility will be. In addition, as in equation 4.2, its expected value
equals the ratio of the total number of opportunities to the total number of oppor-
tunity seekers. Hence, zones in which the residents are on average accessibility-
paar are easily distinguished from zones in which the residents are on average
accessibility-rich. This general accessibility measure depicts an overall picture of
the emerging geography of opportunity, by defining the relative position of each
zone.
The Changing Geography of Opportunity 57
With properly constructed accessibility measures, one can think through more
clearly how spatial technologies - transportation and telecommunications - may
separately and jointly reconfigure spatial relationships in a metropolitan area. One
can effectively explore future changes in the geography of opportunity and ad-
dress some 'what-if' questions systematically and rigorously.
The automobile has been the most important technological force in shaping
American metropolitan areas in the twentieth century. It has fundamentally
changed the geography of opportunity. Using equation 4.1 described in the previ-
ous section, spatial and social effects of the automobile can be characterized in
terms of increases and decreases in accessibility. The spatial effects of the auto-
mobile are largely attributed to four basic changes brought directly or indirectly by
this great invention: speed of access, spatial distribution of jobs and services, spa-
tial distribution of people, and industrial restructuring.
Speed of Access. The automobile has provided the great majority of households
with a fast and flexible means of transportation for daily activities. This great
benefit, however, does not go to the relatively small percentage ofhouseholds that
cannot travel by car due to financial, physical, or other constraints. These people
are still dependent on public transportation. Unfortunately, increased usage of per-
sonal motor vehicles has often been associated with decreased quality of service of
public transportation. Therefore, the automobile has caused a widening gap among
people's ability to overcome spatial separation. It has reduced drivers' zone-to-
zone travel times, but has often increased transit passengers' zone-to-zone travel
times. Everything else being equal, the accessibility for the former would increase,
whereas the accessibility for the latter would decrease.
Spatial Distribution of Jobs and Services. The automobile has enabled urban
employment to decentralize, and has therefore facilitated a spatially dispersed pat-
tern of metropolitan growth. Consequently, a high percentage of private enter-
prises and govemmental agencies are now located in suburban areas not served by
public transit. In the accessibility measure, this change is depicted by reduced op-
portunities in zones located within or near the central city. In the meantime, zones
that are difficult to access by public transit, and therefore have long transit travel
times from other zones, gain a large share of the redistributed opportunities. As-
suming all other factors were constant, accessibility would overall decrease in the
center but increase at the periphery. The greatest decrease in accessibility would
occur for transit-dependent groups living in the central city.
58 Q. Shen
Spatial Distribution of People. The automobile has also effected the decentrali-
zation of urban population. In particular, people who own a private automobile
have gained more freedom in choosing horne and work locations, as weIl as the
freedom to trade accessibility for better housing. The transit-dependent population
groups, however, are confined to residential and employment locations connected
to the public transportation network. For low-income minorities subject to dis-
crimination in the housing and labor markets, location choices are even more lim-
ited. These changes can be properly described in terms of accessibility only if
measurement is made separately for the privileged and the disadvantaged. For the
former, population decentralization has partly offset employment decentralization,
and therefore has reduced the magnitude of accessibility loss in the center and
accessibility gain at the periphery. For the disadvantaged groups, such offsetting
effect is more limited.
Speed of Access. For people who use these new technologies to access opportuni-
ties, the saving in time spent on travel can be enormous. For example, if workers
telecommute two days every week, they will reduce their average daily commut-
ing time by forty percent. Since there are still many households in the United
States who do not have a telephone, it is most likely that a considerable percentage
of the population will not have adequate telecommunications capabilities in the
near future. Therefore, there will be further widening in the gap among people in
terms of ability to overcome spatial separation. Travel time differentials across
population groups will increase. Everything else being equal, the level of accessi-
bility for people who have telecommunications capabilities will increase, whereas
the level of accessibility for the others will decrease relatively.
who are equipped with both automobiles and telecommunications. However, this
group will be relatively small compared to the auto-driving population. Some re-
searchers reported recently that only about a third of employees in the United
States are potential telecommuters (Salomon and Mokhtarian 1997). But overall,
the existing pattern of population decentralization will be reinforced. Therefore,
the offsetting effect, which applies mainly to the economically and socially privi-
leged group, will also occur: population decentralization will reduce the magni-
tude of loss of accessibility in the center and gain of accessibility at the periphery.
The Boston Metropolitan Area, which covers more than two thousand square
miles of land and aeeommodates more than 4 million people, was seleeted for a
ease study. The ease study was foeused on employment aeeessibility beeause
eommuting and teleeommuting data were more available in eomparison with data
for other types of travel or teleeommunieation usage. It eonsisted of two parts. The
first part examined existing patterns of variations of employment aeeessibility
aeross geographie loeations and travel modes. It was an empirieal analysis of 1990
data using equations 4.1 and 4.2. The seeond part explored future ehanges in em-
ployment aeeessibility eaused by a large-seale deployment ofteleeommunieations.
It was a simulation ofsome future scenario using equations 4.5, 4.6, 4.7, and 4.8.
Eaeh part was implemented with a program written in the C language.
This analysis was based on demographie and employment data originating from
the 1990 Census and peak-hour travel time matriees for auto and transit obtained
from the Central Transportation Planning Staff (CTPS) in Boston. The metropoli-
tan area was deseribed by data assoeiated with 787 traffie analysis zones (TAZs).
The measurement of aeeessibility was made with alternative impedanee funetions
- the exponential funetion and several travel-time-threshold funetions, which gen-
erate similar results. The f01l0wing findings were obtained eonsistently, no matter
whether aeeessibility was measured for a1l workers or only for low-skilled work-
ers:
First, the eentral loeation of the inner city still gives the residents advantage in
aeeessing spatially distributed jobs. For workers who use a given transportation
mode, those living near the eentral business distriet (CBD) have relatively higher
employment aeeessibility. In general, workers would not inerease their employ-
ment aeeessibility by moving to the suburbs.
Seeond, although eentral loeation of residenee offers an advantage, auto owner-
ship is the key determinant. More than half of the TAZs are accessibility-rich for
workers eommuting by auto, whereas only a very small proportion of the TAZs
are accessibility-rich for those eommuting by publie transit. In fact, most of the
62 Q. Shen
workers who live in the suburbs and cOinmute by auto have higher employment
accessibility than those who live in the central city but commute by public trans-
portation.
Third, for a large proportion of the TAZs located in the poor neighborhoods of
the central city, the general employment accessibility is rather low because the
level of auto ownership is low. The location advantage of these TAZs is in most
cases more than offset by the low level of auto ownership.
Therefore, in a metropolitan area that is shaped by the automobile, geographic
location is of decreased significance in defining spatial relationships, whereas
travel mode has become critical. Regardless where they live, transit-dependent
groups are faced with a major obstacle in accessing job opportunities and social
services. This conclusion is consistent with the empirical finding that the lack of
auto ownership significantly reduces the chance for low-skilled workers to find
and keep jobs (Ong 1996). It is also broadly consistent with the widely accepted
view that changes in the geography of opportunity have aggravated the economic
and social problems ofthe central city (Kasarda 1995, Teitz 1997, Wilson 1996).
However, it provides a new insight by showing clearly that the problem is not 10-
cation per se, but the increasing spatial barrier for a high percentage of low-skilled
workers who are dependent on the limited service of public transportation. It sug-
gests that planning and policy making should focus on interaction instead of loca-
tion, because even if there were better residential locations for the poor, the
location advantage would be very limited, compared with the disadvantage associ-
ated with transit dependence.
Accessibility Change
o
0
<-0.19
-0.19 - -0.14
0 -0.13 - -0.08 o 5 10 15 20 Miles
-0.07 - -0.02
Figure 4.1. Decrease in employment accessibility for workers who travel by auto and who
are not potential telecommuters.
64 Q. Shen
Accessibility Change
<0.09
0.09 - 0.12
0.13-0.16
0.17 - 0.20 AN
o 5 10 15 20 Miles
Figure 4.2. Increase in employment accessibility for workers who travel by auto and who
are potential telecommuters.
Not surprisingly, the first finding of this part of the study is that for workers
who travel by a given transportation mode, their employment accessibility will
decrease if they do not have telecommuting capabilities. Of course, the effect is
generally the opposite for those who have such capabilities.
The Changing Geography of Opportunity 65
The second finding, which is more intriguing, is that the magnitude of accessi-
bility change is related to the pre-existing geographical configuration of the met-
ropolitan area. Figure 4.1 depicts the pattern of accessibility decrease for workers
who commute by auto and who are not potential telecommuters. 4 The amount of
loss is positively correlated with the pre-existing level of accessibility, as the
greatest losses occur in the central-city area and the smallest at the periphery.5
Figure 4.2 shows the pattern of accessibility increase for workers who commute
by auto and who are potential telecommuters. It also exhibits a distinctive, al-
though more complicated, spatial pattern. In this case, the greatest gains occur in a
suburban ring and the smallest gains both at the center and on the edge of the met-
ropolitan area. Similar relationships between the magnitude of accessibility
change and the pre-existing geography are observed in the results for workers who
commute by public transit. The same findings can also be obtained from Figure
4.3 and Figure 4.4. These scatter charts show the relationships between the magni-
tude of accessibility change and the pre-existing level of accessibility for different
categories ofworkers classified by travel mode and telecommuting capability.
0.30
8, 0.20
c:::
.!
(J
0.10
0.00
~ -0.10
.a
'iij -0.20
j -0.30 1
-0.40
-0.50 -'------------..,.---.,------....;...----~
0.00 0.50 1.00 1.50 2.00 2.50 3.00 3.50 4.00
Pre~xisting Level of Accessibility
i. W~hout Telecommuting Capability With Telecommuting Capability I
Figure 4.3. Relationship between the change and the pre-existing level of employment
accessibility for workers who travel by auto.
4 In Figure 4.1, Figure 4.2 , and Figure 4.5, the equal interval scheme is used to c1assify the
zones into four groups based on the magnitude of accessibility change.
5 Readers who are interested in seeing maps of existing levels of employment accessibility
for workers who travel by different modes in the Boston Metropolitan Area are referred to
an earlier paper by this author (Shen 1998b).
66 Q. Shen
0.15 r - - - - - - - - - - - - - - - - - - - - - - - - - - ,
&
c:
0.10
.! 0.05
c..> 0.00
~ -0.05 C:~~iiI...~ . . . c:---
:s -0.10
'<~
-0.15
Ö -0.20 +--------~-
-0.25 .L...-_ _ _ _--+_ _ _ _ _ _ __ _ _ _ _ _ _ __ ~
Figure 4.4. Relationship between the change and the pre-existing level of employment
accessibility for workers who travel by public transit.
The third finding is that teleeommuting inereases the overall level of employ-
ment aeeessibility for zones where there are higher pereentages of wealthier
households, but reduees the overall level of employment aeeessibility for zones
where poorer households predominate. The primary reason is that low-ineome,
low-skilled workers are less likely to be potential teleeommuters. This outcome is
displayed in Figure 4.5. Geographically, zones that show a substantial decline in
the general level of accessibility are highly eoneentrated in the eentral city.
The fourth finding is that the magnitude of losses and gains in the general level
of employment aeeessibility is usually quite modest. The largest pereentage
change is only 15 percent. Approximately three-quarters of the zones either de-
erease or inerease their general level of aceessibility by less than 5 percent. There-
fore, if teleeommunications only partially substitute transportation, as it is
assumed in this study, the spatial strueture of the metropolitan area will eertainly
not be dissolved.Overall, the deployment of telecommunieations eauses the loea-
tion advantage of the urban center to deerease. In fact, the importanee of geo-
graphie loeation is generally redueed. In that sense, advaneed spatial teehnologies
lead to loeation equalization. This proeess is aeeompanied by further polarization
in aceessibility along the social dimension. The most distinetive social eonse-
quenee is that the wealthier and more edueated residents who ean use the ad-
vaneed spatial teehnologies beeome better off, whereas the poorer and less
edueated residents who eannot use these teehnologies beeome worse-off. The so-
eial distribution of aeeessibility gains and losses is interrelated with, rather than
independent of, the pre-existing geography, whieh supports the argument made by
The Changing Geography of Opportunity 67
Accessibility Change
D <-0.04
Cl -0.04 - 0.00
D 0.01 - 0.05 1\ o 5 10 15 20 Milas
_ 0.06 - 0.10 L..l
N
Figure 4.5. Change in the overall level of employment accessibility when telecommuting
becomes an option for many workers.
Discussion
4.5 Conclusion
This chapter has presented a unified analytical framework for understanding the
spatial and social effects of both transportation and telecommunication technolo-
gies. This analytical framework characterizes these effects by measuring changes
in accessibility. Its key component is a set of accessibility measures that incorpo-
rate not only the traditional land use and transportation components but also the
telecommunications component. Travel impedance is specified for the telecom-
munications component on the basis of empirical evidence that shows partial sub-
stitution of these technologies for travel. The resulting framework allows
consistent and meaningful comparison of accessibility levels across geographic
locations, travel mo des, and telecommunications capabilities.
With this unified analytical framework, one can examine more rigorously the
ways in which the automobile and telecommunications interact with location and
social variables in creating a new geography of opportunity in American metro-
politan areas. In particular, the new framework can serve as a useful tool for un-
derstanding the likely consequences of a large-scale deployment of digital
information and communication technologies. Structured explorations were under-
taken to understand how different socioeconomic groups, defined by their avail-
able options of spatial technologies as weIl as by their residential locations, fare in
the emerging geography of opportunity. Effects on communities were also exam-
ined, by aggregating the effects on the individuals. As an effort to go beyond ab-
stract discussion and generalization, the analytical framework was applied to a
case study of employment accessibility in the Boston Metropolitan Area. Existing
spatial configurations were analyzed, and likely future changes were simulated.
Some of the research findings are not surprising. As it is commonly believed,
advanced spatial technologies reinforce polarization among socioeconomic
groups. The level of accessibility generally increases for those who have tele-
communications capabilities, but decreases for those who depend on more tradi-
tional transportation means. Wealthier communities, inc1uding most of those in the
suburbs, are better off because their residents tend to have telecommunications
capabilities. In contrast, poorer communities, especially those in the central city,
are worse off. Unless effective policies are implemented to make new telecommu-
nications a viable option for disadvantaged groups, the resulting spatial and social
effects will undoubtedly aggravate the problems ofthe central city.
There are important new insights generated by this research. One useful insight
is that geographic location is overall of reduced importance in determining the
potential for spatial interactions in the information age. A good location is no
longer an effective substitution for access to advanced spatial technologies. Spatial
relationships are increasingly defined by technological and fundamentally social
factors. This suggests that in order to improve the situations of disadvantaged
groups, planning and policy making should focus on interaction rather than nar-
rowly on location. Another important insight is that the distribution of accessibil-
ity benefits generated by new technologies is dependent on the existing spatial
70 Q. Shen
References
Atkinson, R.D. 1998. Technologieal change and eities. Cityseape, 3: 129-70.
Batty, M. 1996. Aeeess and information: How the new eleetronie media ehanges everything
geographieal but perhaps ehanges nothing! In CoucIelis, H. (ed.) Spatial Teehnologies,
Geographie Information, and the City. Teehnieal Report 96-10, pp. 42-4. National Center
for Geographie Information and Analysis, University of California, Santa Barbara.
Berry, B. 1973. The Human Consequenees ofUrbanization. New York: St. Martin's Press.
Castells, M. 1996. The Rise of the Network Society. Oxford: Blaekwell.
CoucIelis, H. (ed.) 1996. Spatial Teehnologies, Geographie Information, and the City.
Teehnieal Report 96-10, National Center for Geographie Information and Analysis, Uni-
versity of California, Santa Barbara.
The Changing Geography of Opportunity 71
Gillespie, A. and Robins, K. 1989. Geographical inequalities: The spatial bias of new
communications techno10gies. Journal ofCommunication, 39:7-18.
Gottmann, J. 1983. The Coming ofthe Transactional City. Institute ofUrban Studies, Uni-
versity of Maryland, College Park.
Hall, P. 1996. Revisiting the nonplace urban realm: Have we come full circle? International
Planning Studies 1:7-15.
Handy, S. and Mokhtarian, P. 1996. Forecasting te1ecommuting: An exploration ofmeth-
odologies and research needs. Transportation 23: 163-90.
Hoffman, D.L. and Novak, T.P. 1998. Bridging the racial divide on the Internet. Science
280:390-1.
Kasarda,1. 1995. Industrial restructuring and changing location ofjobs. In Farley, R. (ed.)
State ofthe Union: America in the I 990s, Vo/ume 1: Economic Trends. New York: Rus-
sei Sage Foundation, 215-67.
Meier, R.L. 1962. A Communications Theory of Urban Growth. Cambridge MA: MIT
Press.
MitchelI, W.J 1995. City of Bits: Space, Place and the Infobahn. Cambridge MA: MIT
Press.
Mokhtarian, P.L. 1990. A typology ofrelationships between telecommunications and trans-
portation. Transportation Research 24:231-42.
Morris, J.M., Dumble, P., and Wigan, M. 1979. Accessibility indicators for transportation
planning. Transportation Research A. 13:91-109.
Niles, I.S. 1994. Beyond Te/ecommuting. Washington DC: Office ofEnergy Research, US
Department ofEnergy.
Office of Technology Assessment, US Congress 1995. The Technological Reshaping of
Metropolitan America. Washington, DC: US Government Printing Office.
Ong, P. 1996. Work and car ownership among welfare recipients. Social Work Research
20: 255-62.
Salomon, I. 1986. Telecommunications and travel relationships: A review. Transportation
Research A 20:223-38.
Salomon, I. and Mokhtarian, P.L. 1997. Why don't you telecommute? Access No. 10:27-
29.
Shen, Q. 1998a. Location characteristics of inner-city neighborhoods and employment ac-
cessibility oflow-wage workers. Environment and Planning B 25:345-65.
Shen, Q. 1998b. Spatial technologies, accessibility, and the social construction of urban
space. Computers, Environment and Urban Systems 22:447-64.
Teitz, M.B. 1997. American planning in the 1990s: Part 11, the dilemma ofthe cities. Urban
Studies 34:775-95.
U.S. DOT 1993. Transportation Implications ofTelecommuting. Washington DC: US Gov-
ernment Printing Office.
Webber, M.M. 1964. The urban place and the nonplace urban realm. In M. M. Webber,
editor, Explorations into Urban Structure. Philadelphia PA: University of Pennsylvania
Press, 79-153.
Webber, M.M. 1996. Tenacious cities. In Couclelis, H. (ed.) Spatial Technologies, Geo-
graphie Information, and the City. Technical Report 96-10, pp. 214-18. National Center
for Geographie Information and Analysis, University of California, Santa Barbara.
Weibull,1.W. 1976. An axiomatic approach to the measurement of accessibility. Regional
Science and Urban Economics 6:357-79.
72 Q. Shen
Wheeler, J.O., Muller, P., Thrall, G., and Fik, T. 1998. Economic Geography, Third edition.
New York: John Wiley & Sons, Inc.
Wilson, A.G. 1970. Entropy in Urban and Regional Modelling. London: Pion Limited.
Wilson, WJ. 1996. When Work Disappears. New York: Alfred A. Knopf.
5 Space, Time and Sequencing: Substitution
at the Physical / Virtual Interface
5.1 Introduction
The concepts of access and accessibility are used widely within geography and
owe their origin to early quantitative and theoretical work in the 1960s (Bunge
1966, Warntz 1967, Garrison 1960, Kissling 1966). In a spatial sense, accessi-
bility is both a specific and aggregate property. As a specific condition it de-
scribes individual circumstances, i.e., either whether an individual can access a
point, or whether they can access a facility at a point. The basic question is can
A get to XY? As an aggregate property most working definitions seek to provide
a generalized statement of what can be reached. This often involves showing
the relative advantage of one location over another across what is usually pre-
sented as a smooth surface but in reality is usually a derivative from point val-
ues on a network. An accessibility surface often answers the question of how
many Bs can get to place XY, for all XY?
Aceessibility is generally about physieal movement and physieal aeeess. The
simplest spatial aecessibility measures use isotropie space and Euclidean dis-
tance, but most of the definitions used in research employ a model of space de-
rived from the context of loeal transport systems and the realities of travel and
mobility. Cost and time measures of distance, defined through a transport net-
work model, are the most common definitions of separation used (Weibull
1976; 1980; Pirie 1979). For aggregate descriptions of access, the measure of
distance is frequently modified using a non-linear transformation or a threshold
distance value for access. When travel time is used as a distance measure in the
spatially conceptualized view of access, then mobility and accessibility are
clearly twinned concepts (Dijst and Vidakovic 1997). Pooler (1995) and Handy
and Niemeier (1997) provide reviews of spatial accessibility measures.
There are a number of reasons for questioning spatial measures of accessibil-
ity, some of which relate to their failure to acknowledge non-spatial constraints
on access, such as the use of authority to ban entry to places. Another critique is
to note that accessibility, from a process perspective, is more fundamentally
about whether (or how fully, or at what marginal cost) person A can be at loca-
tion XY for a given purpose. The italicized component often requires a facility
constraint to be met (for instance, a golf course to play golf on). It almost al-
ways requires a minimum temporal stay, so that the individual query above is
Space, Time, and Sequencing 75
rewritten as can A get to XY for duration t? The same is true for the aggregate
queries. Inevitably, the duration requirement is specific to the task to be under-
taken, which is itself constrained within a regime of required, and competing,
personal activities. This is in turn mediated by mobility and changes in spaces
defined by parameters such as travel time (Janelle 1968, Chapin 1974, Kwan
1998, Miller, Wu, and Ming 1999).
This latter, dynamic conceptualization of space and accessibility has major
implications. First, the notion of a generic, purpose-independent accessibility
surface or definition of accessibility becomes untenable. Access varies depend-
ing on the process being considered (the conditions that a particular activity
requires to be met: for instance timing or minimum duration). Second, space
and time (whether modified by transport or not) are no longer an independent
metric for directly mediating accessibility, since the criteria by which a place is
deemed accessible may be predominantly constrained by scheduling options.
These are themselves created by the available options for scheduling other ac-
tivities (which are in turn constrained by different, process-specific time and
space issues). In asense, scheduling is itself a spatial agent in modifying acces-
sibility for the individual, since the ability to interact with others and make use
of facilities or services is conditional upon the existence of, or the ability of
individuals to create, appropriate sequences of time slots to perform activities.
The intermeshing of these demands throughout the day already creates quite
distinct diurnal geographies in the city (Janelle, Klinkenberg, and Goodchild
1998).
Third, in terms of virtual technologies, we see a new partitioning between
what activities can be undertaken virtually through communications media and
what can only be done through physical presence. This is only in part a re-
classification of entire activities into the virtual sphere (see Chapter 14 by
Kwan and Chapter 17 by Occelli). We also see the emergence of a continuum
of optional modes for undertaking specific activities, which allow choice-driven
substitution of virtual for physical presence, or vice versa. Central to this parti-
tioning and choice are issues related to the economics of delivery and the (ac-
tual or perceived) quality of the virtual versus physical experience.
We can expect that these trade-offs vary greatly between different activities
and individuals. In either event, virtual technologies exercise two influences on
issues of access. On the one hand, virtual technologies generally permit the ac-
tor greater flexibility in allowing the re-grouping of activities in a schedule, and
so influence the possible schedule choices for the remaining activities. On the
other hand, physical access to virtual technologies may not be spatially or indi-
vidually ubiquitous (see Chapter 9 by Harvey and Macnab and Chapter 10 by
Moss and Townsend; and see Salomon 1988). Thus, two geographies of access
emerge within any individual's schedule. One is a geography based on the (in-
ternally aspatial) activity sequences of web-based contacts (within which time is
the significant dimension to juggle), and the other is based on the needs of the
physical-presence activities. Of course, these two geographies intermesh, and
over time the nature and interaction of both is changing as technology changes
access to, and capabilities of, the Web.
76 P. Forer and O. Huisman
This chapter seeks to explore these ideas through considering a particular do-
main, in this case that of University students. University education is an excel-
lent context in which to consider the application of new technologies. Student
life styles are currently heavily structured by scheduled teaching situations such
as lectures, laboratories and tutorials. New IT is providing a range of options
that offer trade-offs in the way that these traditionallearning experiences can be
delivered (Forer 1998), This will impact equally on individual lifelines, and
city-wide patterns of accessibility.
The next section juxtaposes issues of evolving student lifestyles with a cellu-
lar space-time approach to modeling individual accessibility (Forer 1993; 1995;
1997; Huisman, Forer, and Albrecht 1997).
Much of the empirie work on which the examples within this paper are based
derives from a study of student life styles at the University of Auckland. The
city of Auckland has over one million inhabitants, and, as in many larger cities,
students must cope with commuting and parking restrietions to study at a cen-
tral site, such as the University of Auckland's main campus, which hosts over
20,000 students. They must also juggle with the unallocated time in their day so
that they are able to sustain paid work, either in the central area or elsewhere.
Further details of these contextual issues can be found in Boswell (1995) and
Forer (1998).
Student lifelines are well suited for analysis because it is possible to identify
major components oftheir necessary daily activities from known course choices
and their related formal teaching timetables. Having defined the suburb a no-
tional student may live in during term time, we can add to this the location of
study and the formal components of their study timetables to define the major
marker episodes in their working day. (i.e., specified activities of which the
location and starting and ending times are known and largely defined (Lennt0rp
1976, Hägerstrand 1982)). It is also possible to use some simplifying assump-
tions to move from records of location and timetable to the creation of a no-
tional activity schedule for domestic chores, sleeping and study for the whole
day, with associated space-time coordinates. This is discussed in Huisman and
Forer (1998a). These mandatory episodes can in turn form the basis for analysis
of individual accessibility to activities in the remaining periods of discretionary
time.
Space, Time, and Sequencing 77
It is argued elsewhere (Forer 1997, and see Chapter 16 by Hanson) that Häger-
strand's conceptualization of lifelines achieved limited deployment because of
the difficulty of making paths and prisms easily operational. These difficulties
related to all facets of generating and modeling activity-based data, for both
lifelines and prisms. In recent years a number of authors have noted the poten-
tial of GIS to cope with two key components in understanding space-time ac-
cessibility: the locational data of individuals, and the issues of movement
through transport networks (Miller 1991, Mey and ter Heide 1997, Löytönen
1998). However, implementations of systems to model lifelines within the ge-
ometry of the space-time aquarium remained relatively few. In 1993 one of the
authors developed a framework for implementing lifelines and prisms in a cel-
lular way, using a combination of raster and network methods to populate a
three-dimensional matrix with binary values denoting absence or potential pres-
ence (Forer 1993; 1995; 1997). This work also proposed defining other actual
or potential space-time objects, for instance an open library, in the same way
within the cellular framework. Figure 5.1 illustrates the kind of representation
developed, as weH as an example of how object types could be combined to
answer spatio-temporal queries through Boolean operations on object masks.
Figure 5.1. An exploded view of three binary masks for a rectangular area over five
time periods. The left-hand mask displays an individual's prism. The central mask repre-
sents three facilities that are open at different times. The right hand mask is a mask ofthe
other two derived from a simple AND operation, and represents the times that the indi-
vidual can be at the facilities while they are open. (Source: Forer 1993).
two basic masks, and a derived accessibility measure of the library access op-
tions in space and time for that individual. The major limitation in deriving such
masks remains the ability to move from coping with clearly defined marker
events to coping with activity schedules where only a desired activity mix,
rather than precise event timing and location, are known.
The construction of prisms is normally based around gaps between known
marker events. For the present implementation, standard allocation procedures
and network data structures are used, in normal and reverse modes, to identify
locations accessible to individuals given a specific travel time from a marker
event and current mobility conditions for the individual. In some circumstances
raster processing is also used to identify travel time across open spaces, such as
large parks. These data are then used to populate a three dimensional binary
matrix. The process for static objects such as libraries is much simpler.
A matrix of space-time object types was subsequently proposed (Huisman,
Forer, and Albrecht 1997), which has been extended to include virtual events
and activities (Table 5.1). Intermittent events are defined by their spatial extent
and location, and the periods of time for which they are accessible, i.e., their
opening or functioning hours. Continuous events represent those objects and
events that exist at any point in time (such as the actual or possible location of
an individual). In the virtual realm these events and activities represent loca-
tions in virtual space (servers or websites) that can be physically (though re-
motely) accessed within temporal constraints.
CU TOMI ED
I DIVID · ALAND
AGGREGATE
VOL ME
POTENTIAL
I TERACTIO
•
RFACE
.
APPU! TIO
DOMAI
L
"-------------"
c=::::>
c: c::>
=> 0 /
AGE TS
A
t ,. T -i
CON TRAlNT
~
I TI
"
SPATlAL
TR CTUREAND
PROPERTIES
Figure 5.2. Extending time-geographie eoneepts. Source: Huisman and Forer (1998a)
from analysis of 100 students. This surface shows the number of discretionary
people/minutes that can be spent at a location by the target group between the
ho urs of 12 no on and 2:00 p.m. of a given day. Essentially it is one measure of
the accessibility of each point to the student body, or of possible student interac-
tion over space during this time period.
In general, then, masks can be derived to show both actual presence (lifelines)
and potential presence (prisms) in a cellular form. These can be combined in
various ways to answer queries and to yield aggregate patterns of accessibility.
The main required ingredient is an ability to specity the space-time location of
marker events. The next section briefly discusses the challenge of applying this
in an operational context where marker events may need to be derived, and
leads on to consider how virtual meetings could be modeled and integrated into
models of accessibility.
Figure 5.3. Potential access and interaction in Auckland. Accessibility surface showing
the number of personlminutes that can be spent at a given location between the hours of
12 noon and 2 p.m. by the sampIe of 100 students. Tbe height of each grid cell is propor-
tional to the number ofperson-minutes for which it can be occupied.
82 P. Forer and O. Huisman
harder to predict, as the relationship between doing an activity and the need for
a physical presence will be determined by individual deeisions on the trade-offs
of using a virtual option.
If we foeus back on aceessibility as we have treated it in this paper, then de-
fining spaee-time accessibility becomes much harder because the marker events
that define diseretionary periods for access are more difficult to formalize, be-
cause at an individual level the real or virtual option could be seleeted. The real
option is tightly location constrained, the latter is either loosely or minimally
constrained. Furthermore, the link between a location and a function may be-
eome much less deterministic. Real activities frequently rely on the nature of
the plaee that hosts them; virtual activities largely import their own cyberplace
to wherever they are being undertaken. Presence at horne may mean university
work is being done, for instanee, as may presenee in a Cyber-eafe in a loeal
mall. In general, we can expect that even a spaee-time accessibility measure
will become a more generalized representation of aggregate aecess, because of
the mutant nature of the possible virtual/real substitutions. We ean also expeet
that to deploy the space-time framework to understand the operation of vir-
tuallreal systems will require a much greater understanding of the aetivity proc-
esses involved, both in isolation and in the specific contexts ofthe individuals.
In spite of these eaveats, it is worth speculating on how we might view the
impact of a restrieted version of virtualized education as diseussed previously.
An implementation of an aceessibility surfaee for one hundred student lifelines
is described below. Issues of access to cyberspace are addressed in a second,
shorter section. The section also looks at how some of these substitutions eould
be modeled in the cellular space-time model deseribed earlier. It identifies is-
sues in inc1uding virtualized activities in the model, notes the centrality of
scheduling in generating new aetivity patterns in these eireumstanees and dem-
onstrates the impact of virtualizing courses under simple assumptions of sched-
ule ehoices.
This analysis works on the known timetables of one hundred notional eom-
merce students in 1996. Their lifelines were developed, and general accessibil-
ity surfaees deriving from their diseretionary time were produced. This process
is described in Huisman and Forer (l998a). In the analysis below, the following
assumptions have been made in respect of modified learning delivery using
virtual technologies:
The students have then been allocated new schedules based on these changes
and their lifelines and prisms have been calculated. From this, new aggregate
surfaces of student accessibility have been derived (Figure 5.4). The left colurnn
shows the original access surface, the right colurnn the surfaces resulting from
physical/virtual substitution.
The analysis largely shows that discretionary time is increased, which should
free up student choice considerably and allow easier access to paid work or
other activities. It also shows the spatial pattern ofthis, and demonstrates a sim-
ple way to identify changes that might spring from a given change in learning
delivery. The basic idea could be deployed with more sophistication, for in-
stance by stochastically modeling virtual/real substitutions for particular activi-
ties. This could be traced through their resultant impact on the overall
individual mask and on the aggregation of masks.
Access to Terminals
A significant tenet of activity modeling in the past has been a relatively tight
mapping of activity to location. In many cases the mapping has been one-to-
one: for instance University classes are only at the University campus. In others
a one-to-many relationship has held, as with shoe shopping being possible at
severallocations. Occasionally one location has supported many kinds of activi-
ties, but this is sometimes due to a scale issue, as when a Mall is treated as a
single entity but supports multiple kinds of shopping activity.
Virtual technologies are unlikely ever to be ubiquitously available, in the
sense of there being no restriction on where high-speed links could be made, or
where high quality screens were available. In the short term (a few techno-
nomads aside), getting on to the Web for virtual services will remain tied to
specific nodes. While these will proliferate very quickly, access to them will
remain restricted by a range of constraints, notably the fact that most will be
within private property and subject to log-on restrictions. In the short term the
nodes into cyber-space are likely to be quite specific. For individuals these
could be horne links, links from educational institutions, links from cyber-cafes
or other pay-for-access locations, or links from work. Some of these locations
will be more or less suitable for certain kinds of virtual links. Surfing or e-
mailing a help desk could be done from a cyber-cafe, but undertaking a de-
manding laboratory exercise or downloading video might require a fast link and
a quiet work area. From a modeling point of view, locating the ports to cyber-
space would be simple, and these possible activity points could be used as part
of a student's lifeline and mask. Once again, however, the specific nature of
particular ports brings us back to a detailed consideration of the student's con-
straints and activity needs.
86 P. Forer and O. Huisman
After Substitution
12:00 noon
4:00 p.m.
BEFORE
- ..
AFTER
Figure 5.4. Potential impacts of physical/virtual substitution. The surfaces above illus-
trate the impact of physical/virtual substitution on accessibility patterns at different times
of the day. Each surface illustrates a ten-minute time interval. These figures cover the
same extent as Figure 5.3. On ce again, height represents the possible duration of pres-
ence (physical access). Lightness is used here to illustrate the results better. Physical
access values above range from 0 (flat) to 1000 personlminutes (highest) for the sampIe
of 100 students.
The discussion above has sought to explore the properties of accessibility that
can be derived from Hägerstrand's dimensional view ofhuman activity. Such a
view is based on atomistic assumptions about an individual's ability to be at a
point (or a point' s ability to be visited), and an aggregation of individual proper-
ties into systematic patterns. It yields a more specific view of accessibility that
is inevitably more closely linked to process and to the scheduling and sequenc-
ing of activities. This is at the same time more powerful but less general than
the traditional spatial formulations of accessibility.
One advantage of this approach is that it implicitly embodies questions that
arise when cyber-activities come into consideration, in that, in general, such
activities are likely to be embedded in a wider process where the virtual/real
trade-off is decided. Many activities, and certainly education, will retain a mix
of virtual and real, where the real component is significant and will retain the
space-time constraints of physical accessibility. The issue in determining how
virtual/real substitution takes place is one that is built around scheduling of the
physical activities within a much less bounded set ofvirtual activities.
In one respect, virtualizing activities throws the entire issue of accessibility
open, in that the one spatial constraint remains the constraint on accessing a
terminal. In another it simply redefines the scheduling task, which as we noted
is in fact a key aspect of defining the marker episodes and thus the residual ar-
eas of discretionary use of time. Aspects of process may define what virtual
activities are best aggregated and best undertaken, and these will then need to
be integrated into a wider context of overall activities. When the necessary vir-
tual and physical markers have been established, a rather different set of acces-
sibility measurements will emerge for the student's discretionary time.
For universities, and for their communities, these changes may have signifi-
cant impacts on the geography of tertiary leaming, although that is not the key
issue under discussion here. The indications are that the ideal form for the uni-
versity, and the likely interactions with its hinterland that are possible, will
change significantly as real and virtual accessibility adjusts to new technolo-
gies. More widely, the same opportunities to interchange the virtual and real
experience will be present in many other spatial processes, including shopping,
working, tourism and socializing. In each case we can expect that physical con-
straints and access will remain important, but that there will be a greater empha-
sis on the nature of the specific processes as determinants of substitution and
88 P. Forer and O. Huisman
the process of activity choice and scheduling. As we know more about different
dimensions of cyber-geographies, perhaps through concepts such as Heikkila's
juzzy club (see Chapter 6), we may be able to develop better ideas on modeling
the two separate geographies (virtual and real) at a more sophisticated level, and
eventually achieve a more satistying integration of the real and virtual compo-
nents of accessibility.
References
Boswell, S. 1995. The Splintering of Student Geographies: Restrueturing of the Tertiary
Edueation Seetor in New Zealand. Unpublished M.A. Thesis. University of Auekland,
New Zealand.
Bunge, W. 1966. Theoretieal geography (2 nd Ed.). Lund Studies in Geography Series C
# 1. Lund: Geerlup.
Chapin, F.S. 1974. Human Aetivity Patterns in the City: What People Do in Time and
Spaee. New York: John Wiley.
Dijst, M., and Vidakovie, V. 1997. Individual action spaee in the city. In Ettema, D. and
Timmermans, H. (eds.). Aetivity Based Approaehes to Travel Analysis. London: El-
sevier Press, 117-34.
Egenhofer, M.J. and Golledge, R.G. (eds.) 1997. Spatial and Temporal Reasoning in
Geographie Information Systems. New York: Oxford University Press.
Forer, P.C. 1993. Geometrie approaehes to the nexus oftime, spaee, and mieroproeess:
Implementing a praetieal model for mundane socio-spatial systems. Paper presented to
Initiative 21, Symposium. Lake Arrowhead, U.S.A.: September.
Forer, P.C. 1995. An implementation ofa diserete 3-d spaee time model ofurban aeees-
sibility. Proeeedings ofthe 1993 Conferenee ofthe New Zealand Geographieal Soci-
ety. Wellington: NZGS, 51-7.
Forer, P.C. 1997. Geometrie approaehes to the nexus of time, spaee, and mieroproeess:
Implementing a praetieal model for mundane soeio-spatial systems. In Egenhofer, M.J.
and Golledge, R.G. (eds.) Spatial and Temporal Reasoning in Geographie Information
Systems. New York: Oxford University Press.
Forer, P.C. 1998. Cyberia and the premature death of distanee. In Peters, M.A. (ed.)
Virtual Teehnologies in Edueation. Wellington: Dunmore Press, 146-70.
Garrison, W.L. 1960. Conneetivity of the inter-state highway system. Papers of the Re-
gional Seienee Association 10: 121-37.
George, G. 1997. Student misery should depress us all. New Zealand Herald, 80etober
A2.
Hägerstrand, T. 1970. What about people in regional seience? Papers of the Regional
Scienee Association 24:7-21.
Hägerstrand, T. 1975. Spaee, time and human eonditions. In Karlqvist A., Lundqvist L
and Sniekars F. (eds.) Dynamic Alloeation of Urban Spaee. Farnborough: Saxon
House,3-14.
Hägerstrand, T. 1982. Diorama, path and projeet. Tijdschrift voor Eeonomisehe en So-
ciale Geografie 73:323-39.
Handy, S.L., and Niemeier, D.A.. 1997. Measuring aeeessibility: an exploration ofissues
and alternatives. Environment and Planning A 29: 1175-94.
Huisman, O. and Forer, P.C. 1998a. Towards a geometrie framework for modeling
spaee-time opportunities and interaction potential. Paper presented at the IGU'98 Con-
Spaee, Time, and Sequencing 89
Eric J. Heikkila
6.1 Introduction
soning with fuzzy sets. This paper is an initial exploration of the implications and
potential of this approach.
The next section reviews fuzzy set theory and fuzzy logic. Several ideas appear
to be of fundamental importance in this regard. One is that fuzzy sets do not obey
the classical 'laws' of non-contradiction and of the excludable middle that apply to
non-fuzzy sets. This leads to characterizations of entropy that appear to have use-
ful potential in the context of urban systems and other systems of accessibility.
The fuzzy set approach also offers a graphical representation of fuzzy clubs that is
highly intuitive and insightful. In the following section, these ideas are translated
and applied more specifically to modeling the impact of information technology
on systems of accessibility. Three fundamental dichotomies are identified: (1)
participation-withdrawal, (2) entropy-order, and (3) strict-loose partitioning of
clubs. The penultimate section of the paper sets out a strategy for measuring these
phenomena empirically.
In classical set theory all sets A are crisply defined, so that the membership in A
of any element Xi of some referential or universal set X is unambiguously defined
with respect to some bivalent membership function mA(x). Each element Xi either
is or is not a member of the crisp set A:
Lotfi A. Zadeh (1965) extended the membership function mA(x) to map from X to
the entire unit interval.
The resulting sets are 'fuzzy' in the sense that the classical yes-no membership
dichotomy now becomes a question of degree or extent. In graphical form, the
extension from crispy to fuzzy sets is represented in Figure 6.1, where in this ex-
The Fuzzy Logic of Accessibility 93
ample the reference set is areal line representation of geographical space. Here,
the urban-rural dichotomy is transformed into a fuzzier notion of urbanity, where
the height of the membership function determines the extent to which a given 10-
cation Xi is a member of the set of all urban locations. In a similar fashion, fuzzy
sets might describe locations that are 'close to work' or 'Iocated in good school
districts'. Neither crispy nor fuzzy sets need be single peaked in general.
Fuzzy Urban
Set m(x)=l
(6.4)
The intersection and union oftwo fuzzy sets A and Bare depicted in Figure 6.2.
The fuzzy set paradigm abolishes two 'laws' of classical set theory. These are the
law 0/ noncontradiction,
AnA'=0 (6.7)
which forbids any overlap between a set A and its complement AC. and the law 0/
the excludable middle,
AuAC=X (6.8)
which states that any set A and its complement A' jointly comprise the universal
or reference set X. In fuzzy set theory, these classical laws only hold for the ex-
treme case that mA(x) E {O, l} 'IIx E X, that is, where the set in question is a crisp
one.
To see that this is so, consider first the law of noncontradiction. By definition, mAc
= l-mA, so
(6.9)
By definition, the measure of elementhood for the null set is zero. That is, m0(xi)
=0 'IIx E X. Thus, the law of noncontradiction can only hold if
The Fuzzy Logic of Accessibility 95
(6.10)
This in turn ean hold only if mA(x) E {O, I} '<:Ix E X, whieh eorresponds to the
classieal definition of a (erisp) set. A similar line of reasoning applies to the law of
the excluded middle. In this case we need
This condition, too, ean only hold if mA(x) E {O, I} '<:Ix E X, and so, as Kosko
(1992) asserts, fuzziness begins where classical bivalent set theory leaves off.
It is useful to introduee two more eoneepts pertaining to fuzzy sets, entropy and
subsethood, before turning to issues of aeeessibility. Entropy measures the une er-
tainty or ambiguity of a system or message. The uneertainty in question need not
be probabilistie; it may be fuzzy in nature. In the eontext of fuzzy sets, entropy is a
measure of the fuzziness of a set, where crisp sets have an entropy measure of
zero, and where the fuzziest set has an entropy measure of unity. Kosko (1992)
answers the question, 'How fuzzy is a fuzzy setT with the following entropy
measure:
(6.12)
(6.13)
The entropy measure E(A) is given as the ratio ofthe eardinalities ofthe overlap
(interseetion) and underlap (union) ofthe set A with its complement AC. From the
preeeding discussion regarding the laws of non-contradietion and of the exeludible
middle, we know that M(AnA C) = 0 and M(AuAC) = I for erispy sets, and so
crispy sets have zero entropy: E(A) = Oll = O. In contrast, maximum fuzziness
oeeurs for sets with mA(x) = 0.5 '<:Ix E X, for in that ease E(A) = 0.5/0.5 = 1. Most
sets williie somewhere between these two extremes.
Subsethood, S(A, B) measures the extent or degree to whieh one set, A, is a sub-
set of another, B. Kosko (1992) operationalizes subsethood as:
Note that in fuzzy set theory, it is possible for a set A to be a subset (to some de-
gree) of another set B that is wholly eontained within it, for in that ease
Club theory, in the tradition ofBuehanan (1965) and Tiebout (1956), examines the
eeonomie rationale by which individuals voluntarily form groups or clusters in
order to derive tangible or intangible benefits through mutual assoeiation. The
club theory literature, whieh is summarized weIl by Comes and Sandler (1986), is
relevant to the formation of social organizations as diverse as municipalities,
health maintenanee organizations, and eountry clubs. The club metaphor is ex-
tended very easily to issues of membership, where the strength or degree of an
individual's membership in, or association with, a club may vary aeeording to
eireumstanee. This notion lends itself very naturally to the formulation of mem-
bership injuzzy clubs:
where mA(x) measures the strength of association that any individual person or
plaee x has with some club A for any individual x belonging to some referential or
universal set X.
Accessibility is a kind of assoeiation, where the more aeeessible a person is to a
plaee (or, more generally, to anode) the stronger is the implied association. As
such, aeeessibility ean be modeled using the same fuzzy set theoretieal approach
deseribed above. Viewed from this perspeetive, aeeessibility is merely one mani-
festation of the more general notion of assoeiation. In a geographie sense, aeeessi-
bility or proximity is an essential measure of the strength or degree of an assoeia-
tion. In a transaetional sense, interaction is an alternative measure of assoeiation.
In a soeio-politieal sense, influence is yet another. Taken together, these ideas
suggest that traditional notions of spatial aeeessibility are misplaeed. Geographie
loeation within an urban area, for example, appears to be seeondary to eonsidera-
tions of 'fuzzy club membership' where those clubs may or may not be geo-
graphieally defined and where club membership need not be binary.
To develop this idea a bit further, we borrow again from Kosko (1992) to de-
velop a geometrie representation of fuzzy clubs. Consider the referenee set X =
{Xl, X2, .. , xn }. Its power set, denoted by 2x, eontains all erisp subsets of X. For
example, in the two-dimensional ease (n=2) whieh is depieted on the left-hand
The Fuzzy Logie of Aeeessibility 97
side of Figure 6.3, 2 x = {0, {XI}, {X2}, X}. These elements of 2x (which are
themselves sets) constitute the vertices of an n-dimensional hypercube such as the
one in Figure 6.3. This is the fuzzy power set of X. Using this representation,
fuzzy sets can be depicted as points within a cube. 2
{X } ________________~ X
~2~
'~.I>~//
B • ../
.. C ..-
o
Figure 6.3. Fuzzy power sets derived from geographie space.
Now consider the stylized map of an urban area on the right-hand side ofFigure
6.3 in which two locations XI and X2 are situated with reference to three urban sub-
centers or nodes A, Band C; and where the distance from XI to A, for example, is
denoted by d AI . Membership ofxl in the fuzzy club denoted by A increases mono-
tonically with decreases in dA I, that is,
(6.17)
For reasons that will become apparent below, it is useful and intuitively reason-
able to impose the following boundary conditions on f(.):
f(O) = 1 (6.18)
lim x --+ '" f (d) = 0 (6.l9)
which (when taken in conjunction with the strict monotonicity in equation 6.17)
state that fuH membership occurs only when there is no distance separating a
member from the center, and that any finite distance yields a positive degree of
membership, however small. These conditions are satisfied by any function with
the familiar negative exponential form
In Figure 6.3 the fuzzy set or fuzzy club A is represented as a point in the fuzzy
power set of X, where its location within the cube indicates the degree of member-
ship of each element of X in A. On the right-hand side of Figure 6.3, XI is located
much closer to A than is Xl. and so its degree of membership in the fuzzy club A is
higher. In contrast, the fuzzy club B is equally accessible from both Xl and X2, and
so it falls on the 45°-line, as does fuzzy club C, although the membership levels of
Xl and X2 in fuzzy club C are less than they are for B, reflecting the fact that C is
less accessible to them than is B.
Three other fuzzy sets are depicted on the left-hand side of Figure 6.3, corre-
sponding to AC, (AnN), and (AuN), whose coordinates in the fuzzy power set of
X are given by
These four points (including the fuzzy set A) are situated symmetrically with
respect to the four vertiees that comprise the (non-fuzzy) power set of X. The fur-
ther these points are from the vertices, the fuzzier the set iso Maximum entropy
oeeurs at the midpoint, whieh happens to be B in this ease, where the four sets
converge so that B = BC= BnB c = BuB c .
Thus far we have established that the language of fuzzy sets as developed by
Zadeh (1965), Kosko (1992) and others aeeommodates rather effortlessly eoneepts
of aeeessibility in geographie or non-geographie eontexts. This is eneouraging, but
the real question is whether the analytieal tools of fuzzy logie ean add to our gen-
eral understanding of aceessibility. We begin this part of the inquiry by probing
more deeply the interpretation of the three fuzzy sets most closely assoeiated with
A, beginning with N. We established already that there is a one-to-one mapping
of fuzzy sets in urban spaee on the right-hand side of Figure 6.3 to points in the
fuzzy power set on the left-hand side of Figure 6.3. What about the inverse map-
ping? Speeifieally, is there a geographie loeation on the right-hand side of Figure
6.3 that eorresponds to the eomplement of the fuzzy club A, and if so, what is its
interpretation?
To answer this question, eonsider the inverse function f -1(.) whieh maps from
elub membership levels to distanee, f -I : m~ d. By taking the inverse of both
sides of equation 6.17 we have
The Fuzzy Logic of Accessibility 99
(6.24)
Building from the definition of the eomplement AC , we may define the 'implieit
eomplementary distanee'
(6.25)
whieh denotes the distanee ofx l from the fuzzy eomplement set AC . From this we
ean infer that N will be loeated somewhere on the set of loeations on the right-
t
hand side of Figure 6.3 that jointly form a eircle of radius (dAI eentered on XI.
We denote this eircle by AI<, where
(6.26)
(6.27)
Thus, the existence and dimensionality of N in the original geographie space de-
pends critically on the dimensionality of the reference set X.
Another assumption from Figure 6.3 that can be relaxed is the explicitly geo-
graphical nature of the original space, with A, Band C representing urban nodes
or subcenters in that example, and with membership or accessibility represented as
a function of geographical distance. Although geographie accessibility is sub-
sumed very easily within the fuzzy club theoretical framework, more abstract no-
tions of association mayaiso be accommodated. For example, the fuzzy clubs A,
B and C in Figure 6.3 may be on-line chat rooms, political power centers or com-
munity.
Theoretical Approach
Vi = I ... n (6.28)
The Fuzzy Logic of Accessibility 101
where T is some other set also belonging to the power set of X. Loosely parti-
tioned sets drawn from the power set of X allow some overlap, but they are still
'partitioned' in the sense that there are some elements of X that are clearly ex-
cluded altogether, and there is another crisp sub set of X that forms the perfect
complement to the set in question. Considering for the moment the case where n is
large, we would say that the set S, whose coordinates in the fuzzy power set of X
are (1, 0, 0 ... 0, 0, I), is more strictly partitioned than T, with coordinates (0, 1,
1.. .1, 1, 0), although neither is strictiy partitioned. Likewise, we say that T is less
strictly partitioned than S. In the extreme case, as T becomes less and less parti-
tioned, it happens that we approach full convergence, corresponding to X itself. In
a social setting, a club that is more strictiy partitioned may cater to narrower inter-
est groups with high degrees of involvement. A religious sect or paramilitary or-
ganization that maintains rigid control over its members, demands a high degree
of commitment from them, and shuns 'outsiders' would be an extreme case of a
strictiy partitioned set. The set of people belonging to the Democratic Party, or
subscribers to the Los Angeles Times, which have clearly defined memberships but
that draw from a large cross-section of the population, are less strictly partitioned.
Now consider the diagonal connecting 0 and X in Figure 6.4. Its midpoint M is
also the midpoint of the fuzzy power set of X and, as discussed earlier, it corre-
sponds to the point of maximum entropy of the system. Conversely, the endpoints
ofthis diagonal, 0 and X, correspond to zero entropy. In the case ofX there is full
convergence and maximum participation. All elements are fully engaged at X. In
the case of 0 there is complete divergence and maximum withdrawal, but there is
no ambiguity or entropy.
X
,, X 13
X 23 X3
,,M
• ,,
A ~
X 12
,, /
XI
/
X2
0
Figure 6.4. Fuzzy power set for n = 3.
To recapitulate, there are three dichotomies that are of interest here. One is the
participation-withdrawal dichotomy, which is measured as distance in any direc-
ti on from 0. Another is the entropy-order dichotomy, which is measured as dis-
102 E.J. Heikkila
tance in any direction from the midpoint M. The third dichotomy is the strictness-
looseness of partitioning, which can be measured as the angle away from the cen-
tral diagonal. Each of these dichotomies is likely to be affected by our collective
movement into the information age. Specifically, as we move from an era of geo-
graphic space to one of cyberspace, a number of changes are likely occurring, in-
cluding:
• movement towards more loosely coupled partitions for the same reasons;
and
Empirical Approach
To implement this approach empirically requires data for some universal reference
set X and the patterns of association each element Xi has with a set of identifiable
clubs; A, B, C .... Z. Obvious candidates for study include (1) daily trip logs of
individuals in an urban area and destination travel zones, (2) telephone records, or
(3) Web browsing logs of a sampie of users. The latter example is used in this
discussion, where destination URLs 3 are the fuzzy clubs to which browsers be-
long, and where frequency and/or duration of visits measures strength of member-
ship. Another advantage of this example is that the World Wide Web is undergo-
ing very fast-paced change, and so the nature of its evolution for the three
fundamental dichotomies is unfolding rapidly and requires shorter time frames. It
mayaiso offer insights for extending Bolton's (1997) notion of places as net-
works.
As before, X represents the universal reference set, in this case a sampIe of n
Web users X = {Xl. X2, ... xn }, where n is large and where the survey records
3 URL refers to the universal record locator that serves as a website's identifier and address
on the World Wide Web.
The Fuzzy Logic of Accessibility 103
cover two distinct periods, t=O ('before') and t=l ('after'). Any website that has
been visited at least once by any user Xi in either period is a fuzzy club. The fuzzy
power set defined by this data sampie is an n-dimensional hypercube and each
website is a fuzzy club located within this hypercube. For this example we define
the membership value mA(xi) for either period as the total time spent in that period
by user Xi at fuzzy club A 4, where all such times are normalized so that the maxi-
mum membership value over both periods is unity. The result is a hypercube simi-
lar to the one depicted in Figure 6.4, but in n dimensions, and with numerous
fuzzy clubs located throughout. The objective of this empirical approach is to de-
termine how the system in question has changed with respect to the three funda-
mental dichotomies: (1) participation-withdrawal, (2) entropy-order, and (3)
looseness-strictness of partitioning. In the context of this example it is reasonable
to ascribe any such changes to the increasing availability of and advances in in-
formation technology, and so this provides a relatively 'pure' case study of how
information technology impacts accessibility.
As noted above, the level of participation in a fuzzy club is measured by the
distance of the club from 0, where the maximum distance is -vn, based on a
Euclidean metric. More specifically, we have
(6.29)
aggregating again over W, the set of all websites, yields an overall level of partici-
pation for period t, Pb and the difference in the level of participation between the
two study periods is given by
(6.30)
As for entropy, we may vary the notation from equation 12 slightly to conform to
current usage:
and then our measure of change in entropy between the two study periods for the
entire system is
(6.32)
And finally, the degree of partitioning represented by any set A is given by the
angle e At formed between the central diagonal and the vector defined by A within
the fuzzy power set ofX in period t. Let AD denote the projection ofthe vector A
onto the central diagonal. Then the degree of partitioning is given by
BAt = COS -1 [M(AD) / M(A)] / (1[/4) for t = 0,1 (6.33)
where the expression has been normalized by (1[/4) so that the maximum degree of
partitioning is one. Then, as with the preceding two cases, we have
(6.34)
Equations 6.30, 6.32 and 6.34 measure the change in participation, entropy and
degree of partitioning, respectively, of the system over the two time periods. As
such, they are useful and practical means of assessing the impact of information
technology on accessibility within a given system. More generally, we can think
about the process by which one fuzzy state is mapped into another. Kosko (1992)
refers to such mappings (from one fuzzy power set to another) as fuzzy systems,
and argues persuasively that neural network techniques, which he terms fuzzy
associative memories (FAM), are especially weil suited to this task. Figure 6.5
illustrates the basic concept. Using fuzzy logic in this manner, neural networks can
be trained computationally to reason with sets rather than with propositions re-
garding questions of accessibility. This idea is not pursued further in this paper,
but it does suggest an intriguing avenue for developing predictive models about
the manner in which accessibility systems unfold.
o
t=O - - - - - - . . t=l
Figure 6.5. Fuzzy associate memories (F AM).
The Fuzzy Logic of Accessibility 105
Fuzzy set theory and fuzzy logic is still relatively new and it is certainly not with-
out its critics, so me of whom argue that the introduction of multivalued member-
ship functions to set theory adds little or nothing that could not be accomplished
through other means. This paper does not seek to resolve that ongoing debate.
However, the evidence presented here suggests quite strongly that the logic of
fuzzy sets has much to offer geographers in their quest to model accessibility in a
geographie and non-geographie context. The following conclusions appear to ap-
ply:
• The language of fuzzy sets can be adapted almost effortlessly to encompass
a generalized notion of accessibility, and has considerable potential for in-
tegrating geographie, social, virtual, and other elements of accessibility or
association;
• This approach also provides a natural extension to the club theory literature
that grew from the Tiebout model, where municipalities or other nodes are
modeled as 'fuzzy clubs';
Acknowledgements
This paper was written in response to the eallfor Participation in the Varenius initiative on
Measuring and Representing Accessibility in the Information Age. I am grateful to the or-
ganizers far prompting this inquiry, to my colleague Niraj Verma for helpful discussions
generally, and to Qisheng Pan far research assistance. I am solely responsible for the pa-
per' s contents.
106 E.J. Heikkila
References
Bolton, R. 1997. 'Place' as 'network': Applications of network theory to local communi-
ties. Paper for Volume in Honor of T.R. Lakshmanan; Department of Economics, Wil-
liams College, Williamstown MA 01267; December.
Buchanan, J.M. 1965. An economic theory ofclubs. Economica. 32:1-14.
Comes, R., and Sandler, T. 1986. The Theory 0/ Externalities, Public Goods, and Club
Goods. Cambridge UK: Cambridge University Press.
Heikkila, EJ. 1996. Are municipalities Tieboutian clubs? Regional Science and Urban
Economics 26:203-26.
Kosko, B. 1992. Neural Networks and Fuzzy Systems: A Dynamical Systems Approach to
Machine Intelligence. Englewood Cliffs NJ: Prentice Hall.
Rescher, N. 1969. Many-valued Logic, New York: McGraw-Hill.
Tiebout, C. 1956. A pure theory of local expenditures. Journal 0/ Political Economy
64:416-24.
Zadeh, L.A. 1965. Fuzzy sets. Information and ControI8:338-53.
7 The E-merging Geography
of the Information Society:
From Accessibility to Adaptability
Daniel Z. Sui
7.1 Introduction
the need to abandon the assumption that communications technology cause cultural,
economic, political and socia! behavior and events and that behavior and events are
consequences of the absence, presence or use of telecommunications hardware. Nei-
ther historical evidence, current events, nor prospects for the future support that as-
sumption.
of course we have access now to a much larger supply of information, but here we
must always keep in mind that the human mind can process at most 50 bytes per sec-
ond - and most of us probably do not do nearly as weil. Few if any persons in our age,
I venture, sufTer from a lack of information - all have stacks of unread papers, unan-
swered e-mail messages, endlessly unopened Internet web pages, memos that are still
in the their brown campus envelopes, videotapes of Public TV programs which we
promise ourselves to watch 'when there is time.' The high priority is no longer in get-
ting the information to us but in selecting and ranking, sorting the duplicative and false
and the irrelevant from the information that we need. Like DNA, most of the informa-
tion is junk.
Human thinking and cognitive abilities are inherently metaphorical, and geogra-
phers have paid increasing attention to the role of metaphors in geographie theo-
ries and models (Sui 2000). Generally speaking, most models in social sciences
are developed according to either a physical/mechanistic metaphor or a biologi-
cal/organic metaphor. The root metaphor behind the various measures of accessi-
bility is a mechanistic one, and the concept of adaptability is motivated by a bio-
logical metaphor. To shift the focus of research from accessibility to adaptability
necessitates not only a change of metaphors but also different policy prescriptions
about the geography of information societies.
From Accessibility to Adaptability 111
1 Unless noted otherwise, data used in this paper come primarily from the following
sources: (I). Telegeography 1999: Global Telecommunication Trafiic Statistics and
Commentary. Gregory Staple (ed.). Washington DC: Telegeography, Inc. (see
http://www.telegeography.com, (2). Network Wizards (www.nw.com), RIPE, Inc.
(www.ripe.net), Web21 (www.web21.com), (3). Clemente, P.c. 1998. State ofthe Net:
The new frontier. New York: McGraw-Hill. Data used in this book are collected by
FIND/SVP 1997 American Internet User Survey (www.cyberdialogue.com. (4). Matrix
Information and Directory Services (MIDS), Inc. (www.mids.org, (5) CyberAtlas, Inc.
(www.cyberatlas.com). and (6) U.S. Department of Commerce, National Telecommuni-
cations and Information Administration (NTIA) 1997. Falling Through the Net /1: New
Data on the Digital Divide (www.ntia.doc.gov/ntiahome/net2/falling.html).
114 D.Z.Sui
Table 7.2. The most wired countries in the world: Top 15 nations by Internet population by
the end of 1998
Nation Internet Users (millions)
United States 76.5
Japan 9.75
United Kingdom 8.10
Germany 7.l4
Canada 6.49
Australia 4.36
France 2.79
Sweden 2.58
Italy 2.14
Spain 1.98
Netherlands 1.96
Taiwan 1.65
China 1.58
Finland 1.57
Norway l.34
Top 15 nations 129.9
Europe 36.02
Worldwide 147.8
Source: CyberAtlas, Inc.
Figure 7.1. Location ofInternet hosts by county, 1997 (Data source: TeleGeography, Inc.)
From Accessibility to Adaptability 115
According to the Computer Industry Almanac, there were more than 147 million
world Internet users at the end of 1998, up from 61 million at the end of 1996.
However, the world Internet population is distributed unevenly. The top 15 na-
tions account for almost 90 percent of the world Internet users, with the Uni ted
States alone accounting 52 percent (Table 7.2). The geographical distribution of
the world's Internet population is consistent with the locational patterns of the
Internet infrastructure (Figure7.1). Internet domains are located predominantly in
the United States (58.1 %) and Europe (24.3%). Both Asia and Latin America ac-
count less than 10 percent of the world's Internet domains. Africa is the least
wired continent, accounting for less than 0.5 percent of world's Internet domains
(Figure 7.2). America's dominance is reflected also in the distribution of the
world's 100 most visited Web sites, among which 94 are U.S. Web sites (Figure
7.2). While the rest of the world claims only 6 percent of these 100 most visited
sites, California alone accounts 40 percent and New York 16 percent.
,
Non-U .S.
6% Oceania Africa
3.1% 0.5%
, ....
••••
Californla
40%
... ~ u.s.
58 .1%
Washington
.
-. .
- ..-;
4%
Figure 7.2. Internet hosts by eontinents and distribution of the 1,000 most popular Web
sites in 1997 (Data souree: TeleGeography, Ine.)
are moving on-line (Table 7.3). Women, the less educated, the less affluent, and
blue-collar workers are increasingly accessing the Internet (Table 7.3). It is esti-
mated that by 2003, Internet users in the United States will reach 207 million. On
average in 1998, Internet users spent 8.2 hours per week on the Net (Table 7.4),
with the majority logging on in the evenings daily between 6 p.m. and 10 p.m.
Since, the Internet will impact significantly on the temporal rhythms of social life,
we will need to pay more attention to 'the urbanization oftime' (Janelle 1993, see
Chapter 9 by Harvey and Macnab).
% Hoosehold On!lIle
c:::::::::J 0.1 - 0.11
c::::J 0.11 - O. 12
0.12 - 0.16
_ 0.16 -0.17
_ 0 .17 -0.19
o
Figure 7.3. Location ofInternet households in the United States. 1997 (Data source: Cy-
berAtlas, Inc.)
Computer Ownershlp
CJ 20.6 - 26 .2
c=J 26.2 -33
33 - 38 .2
_ 38.2 -44 .4
o 44.4 - 62 .6
Figure 7.4. Household computer ownership in the United States, 1997. (Data source: Cy-
berAtlas,Inc.)
From Accessibility to Adaptability 119
At the household level, about 20 percent of U.S. households had 'access to the
Internet by the end of 1997. But it is estimated that this Internet penetration will
reach 70 percent by 2003. With the dramatic drop ofprices ofpersonal computers,
40 percent ofU.S. households own PCs. However, the geographical distribution of
the Internet households and PC ownership in the United States is uneven (Figures.
7.3 and 7.4), with strong concentrations in states where high-tech industry and
R&D activities are located, such as California, Massachusetts, Washington, and
Colorado. Many rural southern states, such as Mississippi, Arkansas, Alabama,
and Kentucky, lag. The disparity in information infrastructure is reflected clearly
in the distribution ofInternet hosts at the state level (Figure 7.5).
Aroosas
Alabama
West VIrginia
South Carolina
Louisiana
KenbJcky
MississipPI
Kansas
Tennessee
Nevada
Montana
Oklahoma
Indiana
S outh Dakota
Florida
N orth Carolina
Georgia
New Jersey
IllinOIS
WYOffi1l1g
OhlO
Hawaii
Alaska
Delaware
Arizona
Texas
Connecttcut
lewa
Missouri
Vermont
Idaho
New Hampshire
Pennsyrvania
Rhode Island
New York
Washington
Wlsconsm
Maine
N orth Dakota
Oregon
Michigan
NebraSka
Maryland
Colorado
Minnesota
Mass achusetts
Califomia
Virt~~
D.G
New Mexico
o
o Ln
o o
Figure 7.5. Distribution ofU.S. Internet hosts per 1,000 people by state, 1997
120 D. Z. Sui
Table 7.5 shows America's most wired cities. Not surprisingly, 72 percent of
adults in the San Francisco Bay area access the Internet regularly, making it the
most wired community in the nation. This relates obviously to the nearby presence
of Silicon Valley and its concentrations of high-tech companies and highly skilled
people. Other most wired cities in the United States tend to be the traditional eco-
nomic, financial, and political centers, or the new booming Sunbelt cities, where
new corporate headquarters tend to be located or relocated. Figure 7.6 shows the
distribution of Internet domains at the zip code level in Texas. There exist, obvi-
ously, heavy concentrations in the four largest Texas cities - Houston, Dal-
laslForth Worth, Austin, and San Antonio. Several zip code areas in Austin claim
the heaviest concentrations of the Internet domains. This map also shows the
growing dispers al to suburban areas of these four big cities, and even to small
towns and less populated areas in rural Texas.
From Accessibility to Adaptability 121
Di stri buti on of
...,...-, .
Interne t Domai ns
in Texas, 1998
Each dot r ~pr e$ents t he
centrol d of a ZI peode area
.· M"'
o
,
l00Mlles
'
• • •• ••
_-----_...J •• •
• ••
· ..d.
••
Figure 7.6. Distribution ofIntemet domains in Texas, 1998 (Data source: MIDS, Inc.)
Although the results are far from conclusive, preliminary findings do provide
hints to the E-merging geography of the information society. Contrary to the sim-
plistic, utopian expectations that telecommunications have sentenced distance to
death and render geography meaningless, the E-merging geography of the infor-
mation age is producing new rounds of unevenness and geographical clustering.
The ubiquitous availability of information has not created (and will not create) a
more egalitarian even distribution. Despite business es that increasingly conduct
commerce without geographical propinquity, the development of the information
122 D. Z. Sui
soeiety is eoaleseing into an inereasingly clustered spatial fonn at global and na-
tional seales, with a rapid dispersal at loeal seales. The Internet-Ied teleeommuni-
eation revolution simultaneously refleets and transfonns the topologies of eapital-
ism, ereating and rapidly reereating nested hierarehies of spaees teehnieally
artieulated in the global information infrastrueture led by the architecture of com-
puter networks. Indeed, as Warf (1997, 9) observed,
far from eliminating variations among places, such systems permit the exploitation 01'
differences between areas with renewed ferocity. The new geography engendered by
the telecommunication revolution dos not entail the obliteration of local uniqueness,
only its transformation and reconfiguration.
Findings ofthis research are consistent with Feldman and Florida's (1994) and
Gertler's (1995) work on spatial diffusion of innovations and on conclusions
reached from a more eonventional political economy approach (Harrison, Gant,
and Kelly 1996, Scott 1996, Storper 1997, Cooke and Morgan 1998). Such uneven
development reveals the tendency of path dependence to lock-in already-
established patterns of regional development (Rosenberg 1982). History has a
strong vote (although not a veto) in determining what geographie patterns emerge.
Exciting theoretical groundbreaking work has been conducted to explain path de-
pendence through the mechanism of increasing returns and positive feedback in
situations of multiple equilibrium (Arthur 1994, Krugman 1997; 1998).
Like it or not, the information age has arrived (at least in the United States, ifnot
in the world) and it is here to stay. The U.S. economy has undergone a major ge-
netic mutation from the production and shipment of tangible goods to the genera-
tion and cireulation of intangible information and knowledge. To better under-
stand the geography of the E-merging information soeiety, we must liberate
ourselves from the tyranny of mechanistic conceptualizations of reality in the Car-
tesian mode. The information society is obviously a complex adaptive nonlinear
network. Conventional conceptualizations of reality, according to deterministic,
reductionistic, ahistorical, and teleological assumptions, may be sufficient for the
industrial age, but are ineapable of eapturing the unpredictable, interconnected,
and constantly evolving new information soeiety.
The central arguments of this chapter are that the importance of access and ac-
cessibility declines as new innovations mature and that information overload is as
harmful as information scarcity. What really matters in the information age is
adaptability, which entails how individuals, organizations, or regions leam from
what they have access to and translate knowledge into productive use. By shifting
focus from accessibility to adaptability, we will not only enhance understanding of
the eomplexity and evolution ofthe information age better, but also bring life back
into geographie research by ridding it ofthe mechanistic metaphor.
The geographical understanding of the information age demands more than just
another twist of accessibility measurements by modifying the various versions of
the Newtonian gravity model. The adaptability perspeetive proposed in this chap-
ter demands a fundamental paradigm shift in geographie research at the semantic
and syntactic, as weIl as pragmatic, levels. At the semantic level, we are dealing
with a fundamentally different kind of society as a result of the telematie revolu-
From Accessibility to Adaptability 125
tion. Many old concepts and theories are no longer applicable and new theories for
understanding this new reality have yet to be developed. At the syntactic level,
developments in science and technology during the later half of 20th century pro-
vided new languages to describe and model various facets of society. These new
theories and concepts, reflected in chaos theory, cellular automata, fractal geogra-
phy, and self-organizing theory, are coalescing rapidly into a non-linear science
that challenges deterministic and linear thinking of the Newtonian tradition. Al-
though we may never be able to eliminate surprises from our models, we can still
hold out the possibility of creating something approximating what Casti (1994)
called a science of the surprising.
As for policy implications, the adaptability perspective emphasizes the cognitive
and learning capabilities of individuals, organizations, and geographie regions
(Senge 1990) instead of the single-minded pursuit to satisfy the insatiable des ire to
have unlimited access to resources and information. The adaptive paradigm seeks
to balance economic efficiency, social equity, and environmental sustainability so
that society and environment can co-evolve in harmony and avoid further trage-
dies for the commons. Philosophically, adaptability shifts the focus from access-
ing without to learning within. From the adaptability perspective, it is far more
important for an individual, an organization, or a geographie region to be inspired
than to be wired. Studies on the geography of information society will eventually
gravitate toward the three fundamental questions that Immanuel Kant (1998)
raised in Critique of Pure Reason. Who are we? What can we do? What should
we do? Research on the information society may fail miserably if we do not step
back to ponder on these more fundamental questions. I strongly believe that a re-
search focus on adaptability may provide better answers to Kant's questions than
would a focus on accessibility. Adaptability implies a change of human con-
sciousness that is essential if the world is ever to be transformed to a sustainable
state.
Acknowledgements
The author would like to thank Don Janelle for substantive and editorial comments that
significantly improved the quality of this chapter; Shawn Dawson and Kristi Rudy at
MIDS, Inc. for helping me compile the Internet domain data at the zip code level for the
state ofTexas; Paul Adams for referring me to TeleGeography'1999; and Fayu Lai for re-
search assistance in data processing.
References
Abler, R.F. 1991. Hardware, software, and brainware: mapping and understanding tele-
communications technologies. In Brunn, S.D. and Leinbach, T.R. (eds.) Collapsing
Spaee and Time: Geographie aspeets of eommunieations and information. New York:
Harpercollins Academic, 31-48.
Arrow, KJ. 1962. The economic implications of leaming by doing. Review of Eeonomie
Studies 29: 155-73.
Arthur, W.B. 1994. Inereasing Returns and Path Dependenee in the Eeonomy. Arm Arbor
MI: The University ofMichigan Press.
126 D. Z. Sui
Hägerstrand, T. 1975. Space, time and human condition. In Karlkvist, A., Lundqvist, L.,
and Snickars, F. (eds.), Dynamic Allocation o/Urban Space, Farnborough: Saxon House,
3-12.
Handy, S.L. and Niemeier, D.A 1997. Measuring accessibility: an exploration ofissues and
alternatives. Environment and Planning A 29: 1175-94.
Hanson, S. 1998. Off the road? reflections on the transportation geography in the informa-
tion age. Journal o/Transport Geography 6:241-9.
Hanssen-Bauer, 1. and Snow, C. 1996. Responding to hypercompetition: the structure and
processes of a regionallearning network organization. Organization Science 7 :413-27.
Harrison, B., Gant, J., and Kelly, M. 1996. Innovative firm behavior and local milieu: ex-
ploring the intersection of agglomeration, industrial organization, and technological
change. Economic Geography 72:233-58.
Harvey, D. 1996. Justice, Nature and the Geography 0/ DifJerence. Cambridge MA:
Blackwell Publishers.
Hodge, D.C. 1997. Accessibility-related issues. Journal o/Transport Geography 5:33-4.
Hodgson, G.M. 1993. Economics and Evolution: Bringing Life Back into Economics. Ann
Arbor MI: University ofMichigan Press.
Holland, 1.H. 1998. Emergence: From Chaos to Order. Reading, MA: Addison-Wesley.
Hughes, T.P. 1983. Networks 0/ Power: Electrification in Western Society, 1880-1930.
Baltimore MD: John Hopkins University Press.
Janelle, D.G. 1993. Urban social behaviour in time and space. In Bourne, L.S, and Ley,
D.F. (eds.) The Changing Socia! Geography o/Canadian Cities Montreal and Kingston:
McGill-Queens University Press, 103-18.
Jin, DJ. and Stough, R.R. 1998. Learning and learning capability in the Fordist and Post-
Fordist age: an integrative framework. Environmental and P!anning A 30: 1255-78.
Kant, I. (Guyer, P., and Wood, AW., trans.) 1998. Critique 0/ Pure Reason. New York:
Cambridge University Press.
Kelly, K. 1998. New Ru!es tor the New Economy: Ten radical strategies tor a connected
world. New York: Viking.
Krugman, P. 1997. How the economy organizes itself in space: A survey of the new eco-
nomic geography. In Arthur, W.B., Durlauf, S.N. and Lane, D.A (eds.) The Economy as
an Evolving Complex Systems 11. Reading MA: Addison-Wesley, 239-62.
Krugman, P. 1998. What's new about the new economic geography? Ox/ord Review 0/
Economic Policy 14:7-17.
Kwan, M.P. 1998. Space-time and integral measures of individual accessibility: A
comparative analysis using a point-based framework. Geographical Analysis 30:191-216.
Leebaert, D. (ed.) 1998. The Future o/the E!ectronic Marketplace. Cambridge MA: The
MIT Press.
Leydesdorff, L.A. and van den Besselaar, P. 1994. Evolutionary Economics and Chaos
Theory: New Directions in Technology Studies. New York: St. Martin's.
Litan, R.E. and Niskanen, W.A. 1998. Going Digital: A guide to Policy in the Digital Age.
Washington DC: Brookings Institutions.
Lundvall, B. and Johnson, B. 1994. The learning economy. Journal o/Industry Studies
1:23-41.
Lynch, A 1999. Thought Contagion : How Belie/ Spreads Through Society. New York:
Basic Books.
128 D. Z. Sui
Margherio, L. 1997. The Emerging Digital Economy. Project report prepared for the De-
partment ofConunerce. Available on-line at http://www.econunerce.gov
Marshali, A. 1961. Prindples 0/ Economics (9th ed., with annotations by C. W. Guille-
baud). London: Macmillan for The Royal Economic Society.
Miller, H.J. 1999. Measuring space-time accessibility benefits within transportation net-
works: Basic theory and computational procedures. Geographical Analysis, 31: 1-26.
Mirowski, P. (ed.).1994. Natural Image in Economic Thought: 'Markets Red in Tooth and
Claw. ' Cambridge: Cambridge University Press.
Mirowski, P. 1989. More Heat Than Light: Economics as Sodal Physics and Physics as
Nature's Economics. New York: Cambridge University Press.
MitchelI, W.J. 1995. City 0/ Bits: Space, Place, and the In/obahn. Cambridge MA: MIT
Press.
Mody, A. 1993. Learning through alliances. Journal 0/ Economic Behavior and Organiza-
tion 20:151-70.
Mokyr, J. 1997. Are we living in the middle of an industrial revolution? Economic Review
82:31-44.
Moore, J.F. 1993. Predators and prey: A new ecology of competition. Harvard Business
Review May/June, 75-86.
Moore, J.F. 1996. The Death 0/ Competition: Leadership and Strategy in the Age 0/ Busi-
ness Ecosystems. New York: Harper Business.
Nelson, R.R. 1995. Recent evolutionary theorizing about economic change. Journal 0/
Economic Literature, 53:48-90.
NUA. 1998. How MaYTY Online?
(http://www.nua.ie/surveys/how_ many_ online/index.html).
Ohmae, K. 1990. The Boderless Wor/d New York: Harper Business.
Ohmae, K. 1995. The End o/the Nation Sate: The Rise o/Regional Economics. New York:
Free Press.
Paton, R. 1994. Computing with Biological Metaphors. New York: Chapman & Hall.
PetchelI, J. 1993. From production systems to learning systems. Environment and Planning
A 15:797-815.
Powell, W. and Smith-Doerr, L. 1994. Networks and economic life. In Smelser, N. and
Swedberg, R. (eds.) The Handbook 0/ Economic Sociology. Princeton NJ: Princeton Uni-
versity Press, 368-402.
Rogers, J.D. 1998. Executive sununary of GVU's WWW User Survey. Available at
http://www.cc.gatech.edulgvuluser_ surveys/survey-1998-04/
Rogerson, P.A. 1996. Estimating the size of social networks. Geographical Analysis 29:50-
63.
Rosenberg, N. 1982. Inside the Black Box: Technology and Economics. Cambridge: Cam-
bridge University Press.
Schindler, P.E. 1998. Horne Internet use breeds depression.
http://www.techweb.com:3030/internetlstoryININ 19980831 SOOO I
Scott, A. 1996. Regional motors ofthe global economy. Futures, 28:391-411.
Senge, P.M. 1990. The Fifih Discipline: The Art 0/ Practice 0/ the Learning Organization.
New York: Doubleday.
Shen, Q. 1998. Location characteristics of inner-city neighborhoods and employment ac-
cessibility oflow-wage workers. Environment and Planning B 25:345-65.
From Accessibility to Adaptability 129
1 Centre for Advanced Spatial Analysis, University College London, 1-19 Torrington Place,
London WClE 6BT, UK. Email: m.batty@ucl.ac.uk
2 Department of Geography, University of Utah, 260 S. Central Campus Drive Room 270,
Salt Lake City UT 84112-9155, USA. Email: harvey.miller@geog.utah.edu
create new spaces and new patterns of human behavior. In short, new definitions
and conceptualizations of accessibility can only be defined by mapping physical
or material space onto virtual or ethereal space, thus defining a nexus of hybrid
space which, we will argue here, represents the appropriate focus for a new geog-
raphy of the information age.
A popular example serves to make our point. The current fascination with the
online bookshop Amazon.com, which is mentioned many times in this book, is
based on the notion we can substitute making a physical visit to the bookshop with
a virtual visit, even engaging in price comparison, reducing the need for several
visits to different places. However although much of our behavior in browsing and
purchasing is removed from the physical realm, ultimately the book needs to be
delivered in its material form and this depends on where it is warehoused and how
it is shipped. In the case of Amazon.com, there are 6 warehouses in the United
States and 2 in Europe, strategically located to minimize physical transfer costs
and to maximize access to population centers, thus reinforcing long established
ideas that location ultimately depends upon physical accessibility (Dodge 1999).
Of course bookshops are probably not the best example as the product itself can
easily be made virtual - it does not need to be material - although food shopping
and other popular kinds of e-commerce reinforce the point.
We can visualize the interpenetration of these two kinds of space, although in
themselves they are much variegated, as an intersection of two worlds. lt is even
possible that there is simply one world, for one cannot exist without the other,
although there is an assumption that the physical world existed prior to the virtual.
In the sense of the virtual world being exclusively cyberspace, this is indeed the
case. However, it is convenient to consider these domains as intersecting but not
being coincident, for this serves to show how geographers (and all other social
scientists, of course) often abstract them as one or the other, thus providing per-
spectives from one viewpoint or the other.
Place ----.
Non-Place ----. SPACE
.....''----------
~
Figure 8.1. Geographie abstraetion of physieal, virtual and hybrid worlds.
Physical, Virtual and Hybrid Information Spaces 135
Although virtual space is aspatial and does not correlate weil with geographie
space, it is c1ear that virtual space and physical space influence each other. Activ-
ity in virtual space can affect activity in physical space and vice-versa. For exam-
pie, virtual interaction can be both a substitute and a complement to physical in-
teraction. An example of the former relationship is when an individual shops
online rather than visiting a retail establishment. An example of the latter case is
when an individual uses the Web to find a new restaurant or plan a vacation.
To date, there is little research on the interactions between activities in virtual
space and physical space (but see, however, Salomon (1986); and Shen (1998)).
There is a tradition within the geographie literature on measuring the interaction
between physical interaction and perceived/experienced geographie space. For
example, Abler (1975) explores the impact of space-adjusting technologies on
human activities in geographie space while Janelle (1968, 1969, 1991) has pio-
neered the concept of time-space convergence to describe the radical impacts of
transportation on spatial relationships. Reginald Golledge and Waldo Tobler have
developed analytical techniques for transforming geographie space based on the
perceived distances and observed interaction patterns (see, e.g., Golledge and
Spector 1978, Tobler and Wineburg 1971, Tobler 1976, 1978), but the usefulness
of these techniques for visualizing hybrid space has not yet been explored.
Identifying significant centers and locations in both the virtual and material
worlds is also an important task for future research. Of equal importance to meas-
uring flows is research into the content of such flows and into the processes that
mediate these flows. Markets are increasingly structured in real time across elec-
tronic networks. This poses a level of complexity on the real world that makes
traditional market analytical techniques untenable. Also important is developing
c1ear notions of the demand for and supply of information, particularly with re-
spect to highly diverse networks where there are already very c1ear distinctions in
terms of usage. What is available and what is required for what purposes are very
different notions that must be identified, not only in relation to new information
spaces but also to how these spaces map onto existing physical spaces. These may
be articulated at various scales from social networks to global markets.
Questions as to the quality of interactions in and hybrid worlds are also central
themes. Despite progress in immersive and virtual reality technologies, it is likely
that qualitative differences between interaction using physical and virtual modes
will persist for some time. Previous research on consumer search behavior c1early
demonstrates that individuals use different modes depending on the type of infor-
mation sought (for a review, see Miller 1993). As virtual and hybrid worlds ex-
pand to encompass more aspects of daily life, it is likely that a complex partition-
138 M. Batty and H. J. Miller
ing of aetivities among these modes will oeeur. The nature of this partitioning is
far from c1ear.
A related theme eoneems the quality of information within physieal, virtual, and
hybrid spaees. Vagueness and fluidity eharaeterize virtual spaees. Apparently
good and bad information spaees ean be eontrasted with good and bad information
within these spaees with no one-to-one eorrespondenee between eaeh. The transi-
tory nature of the digital world eontrasts with the material world where informa-
tion spaees are usually struetured in terms of built form that has a life span with
some permanenee. This foeus on temporality is an issue that serves to test the lim-
its of our debate, reinforeing the long standing idea that time geography and ae-
eessibility in time, as well as or rather than spaee, is of mueh more signifieanee
here than we had hitherto thought.
The c1iehe of the digital world - that networks enable people to interaet with
anyone, anywhere, at any time and in any place - illustrates our erude vision of
the emerging digital world. Here our foeus is mueh more eonsidered with an em-
phasis on how humans interaet with one another in spaee and time, adapting to
aeeess the right amount and the right information in the right time and in the right
place. Harvey and Maenab (Chapter 9) c1early illustrate these issues with their
diseussion of interpersonal temporal aeeessibility at the global seale. The ability to
interaet in real-time may be the eritieal faetor that distinguishes among world eul-
tures rather than traditional notions of geographie separation and determinism.
Another theme that weaves its way through the debate involves the development
of tools and protoeols to enable efficient navigation through information spaee.
This instrumental viewpoint suggests that good geographical metaphors, grounded
in good theory about the information soeiety, should be at the basis of navigation
tools that link behavior to purpose. Mueh visualization work is eoneemed with the
development of better tools. These tools are being developed and tempered by the
various institutional struetures that require them. At the same time, we are begin-
ning to leam more about new information spaees using the very tools designed for
using these spaees in a routine fashion. As we develop tools to explore these new
spaees, these tools are being used in routine ways to navigate them. At the same
time, the existenee of these tools modifies these spaees. This is a kind of relativ-
ism that is rarely highlighted in the material world.
Although there are few indieators of the kinds of strueture and behavior, form and
proeess, that are determining the shape of these new information spaees, there are
Physical, Virtual and Hybrid Infonnation Spaces 139
Figure 8.2. The new geography of economic potential: accessibility to Internet hubs (from
Shiode and Dodge 1998).
Casual evidence also reveals the importance of the local and the global in the
morphology of cyberspace. Any casual examination of your email log will reveal
distance decay around your local site that is obvious enough in that most email
deals with human activity through a virtual medium in a local physical space.
However for academics and increasingly for the public at large logged into the
Web, such local interaction is being supplemented with global interaction, which
binds our social networks together in ways that serve to strengthen the global
economy. New network studies of the small-world problem (Watts and Strogatz
1998) suggest that such occasional global ties increase interaction much more sig-
nificantly than the number of such ties might suggest and, thus, studies of the net
and nets using such ideas appear promising.
For the shape ofthe net, there is enough evidence to show that this too is scaling
- fractal- in that links (through Web sites for example) are scaling in importance,
and follow the classic dendritic pattern that we find in many areas of morphology
from the growth of crystals to the growth of cities and other organisms and or-
ganizations. In Figure 8.3(a), we show a piece of the network diagram of the net
produced by Bill Cheswick and Hai Burch at Bell Labs, which has clear fractal
structure. The particular segment - its nodes and links - is not important per se but
its fractal structure iso In Figure 8.3(b) we show the morphology of a virtual com-
munity in cyberspace - Alphaworld - reported on by Dodge (Chapter 11), which
has also evolved as a fractal around the point that members enter this virtual world
- at ground zero. There is no friction of distance in Alphaworld but its physics is
dictated by who came first and the town has evolved around ground zero and
along easily recognizable radial routes.
Physical, Virtual and Hybrid Information Spaces 141
b
Figure 8.3. Morphologies ofvirtual space: cyberspace. (a) the fractal structure ofthe Inter-
net (http://cm.bell-labs.com/cm/cs/who/ches/map/index.html); (b) the spatial structure ofa
virtual world - Alphaworld (see Dodge, Chapter 11).
142 M. Batty and H. J. Miller
,,' "'1.
"1::..a: .....- ...
...-.-.
.,. 11:
·004'. 0", ....1..'"
-.(.
. I • .' ..r..r •
_.
a._
..... '.
.:::e....... r:..".. :.~
~•• I.~
r
...
I. -. . ••
• r,.
~.....&.II.....:
~.:
". :.:-,. ...._-:.
. it·
.I:.:-" :. .........._.-'r
, _-:r-....- .- a:. ._ _.:.. ••
t..:..:.~ r...... ;"'. .,. .........._...
I. ......:... t •••
.:..' .- a. ..
.•.•
I . '• • • • • •
.):. +" r
0.1~:-
......
'or ~
• - •
0,r-
b
Figure 8.4. Morphologies of real space: Euclidean geographical space. (a) route structure
ofa medium-sized English town; and (d) model ofurban growth based on diffusion-limited
aggregation.
Physical, Virtual and Hybrid Information Spaces 143
We will now attempt to cull all these deliberations into a plan for future research
based on (i) what do we know? and what do we have? (ii) what are appropriate
future research directions? and; (iii) what are appropriate research questions? We
will deal with these in turn. AIthough the various debates that follow in the chap-
ters within this section cover a wide range of issues, we will map out our research
agenda in terms of the initial themes introduced here, stating these as questions
that frame research directions.
We begin by considering 'What Do We Know about Visualizing and Represent-
ing Information Space?' A thorough review of these questions is required. This
might be accomplished through research projects but it is more Iikely to come
from the current generation of researchers, such as those of us writing here, com-
ing to concIusions similar to our own and spontaneously developing such reviews
and statements. There are several themes that might spin-off from such reviews
and we will list these:
These reviews begin to merge into major research questions to define the re-
search frontier, and there are some obvious areas that require research programs:
again we state four of these to provide some sense of where we consider the focus
should be:
• Researehing the Flow and Cost of Information. How flows can be identi-
fied and linked to the emergence of new spaces, which, in turn, map onto ex-
isting market, social, and institutional processes.
• Tools of Cybernavigation. The development of new tools for both exploring
and moving through information spaces that are based on insights into the
emergence of such spaces, the interface between activity in real and virtual
worlds, and developments in human-computer interaction.
• Mapping Activity Spaces. Exploring ways in which existing approaches
within time geography can be informed and extended by network paradigms,
network flow data, and scientific visualization.
• Visualization of Connections between Virtual and Real/Material Geog-
raphies. Providing insights into how information spaces are connected to real
spaces through augmenting existing measures of accessibility and the devel-
opment of new ones.
Much more specific research issues can also be identified, which might drive
forward the research agenda. There is an urgent need for a major initiative in the
collection of network data and its subsequent analysis with respect to the search
for new information spaces. These initiatives could take many forms and we list
four here:
We also need to evaluate the role of existing tools in spatial analysis and to de-
velop new tools relevant to the issues we have identified pertaining to the analysis
Physical, Virtual and Hybrid Infonnation Spaces 145
References
Abler, R 1974. Effects of space-adjusting technologies on the human geography of the
future. In Abler, R, Janelle, D., Philbrick, A, and Sommer, J. (eds.) Human Geography
in a Shrinking World. North Sciutuate MA: Duxbury Press, 35-52.
Batty, M., Dodge, M., Doyle, S. and Smith, A 1998. Modelling virtual environments. In
Longley, P.A, Brooks, S.M., McDonnell, R., and Macmillan, B. (eds.) Geocomputation:
A Prim er. Chichester UK: John Wiley, 139-61.
Beguin, H. and Thisse, l-F. 1979. An axiomatic approach to geographical space. Geo-
graphical Analysis, 11 :325-41.
Caimcross, F. 1997. The Death ofDistance. Cambridge MA: Harvard Business Press.
Castells, M. 1996. The Rise ofthe N etwork Society. Oxford UK: Blackwells.
Clarke, K. 1998. Visualising different geofutures. In P.A Longley, Brooks, S.M., McDon-
nell, R., and Macmillan, B. (eds.) Geocomputation: A Prim er. Chichester, UK: John
Wiley, 119-37.
Dodge, M. 1999. Finding the source ofthe Amazon.com: Examining the truth behind the
hype of the 'earth's biggest bookshop'. A paper presented to the 5th E*Space Confer-
ence, Cape Town, South Africa, July 1999.
Golledge, RG., and Spector, A N. 1978. Comprehending the urban environment: Theory
and practice. Geographical Analysis 10:403-26.
Janelle, D.G. 1968. Central place development in a time-space framework. Professional
Geographer, 20:5-10.
Janelle, D.G. 1969. Spatial organization: A model and concept. Annals ofthe Association
ofAmerican Geographers 59:348-64.
Janelle, D.G. 1991. Global interdependence and its consequences. In Brunn, S.D., and
Leinbach, T.R. (eds.) Collapsing Space and Time. London: Harper-Collins, 49-81.
Markoff, 1 1999. Not a great equalizer after all ? On the Web, as elsewhere, popularity is
self-reinforcing. The New York Times (Monday June 21, C4).
Miller, H.l 1993. Consumer search and retail analysis. Journal ofRetailing, 69:160-92.
Miller, H.l 1999. Measuring space-time accessibility benefits within transportation net-
works: Basic theory and computational procedures. Geographical Analysis, 31:187-212.
MitchelI, W.l 1995. City of Bits: Space, Place and the Infobahn. Cambridge MA: MIT
Press.
146 M. Batty and H. J. Miller
Muller, J-c. 1982. Non-Euc1idean geographie spaces: Mapping functional distances. Geo-
graphical Analysis, 14:189-203.
Negroponte, N. 1995. Being Digital. New York: A. A. Knopf.
Salomon, I. 1986. Telecommunications and travel relationships: A review. Transportation
Research A, 20:223-38.
Shen, Q. 1998. Spatial technologies, accessibility and the social construction of urban
space. Computers, Environment and Urban Systems, 22:447-64.
Shiode, N., and Dodge, M. 1998. Using GIS to analyse the spatial pattern ofthe Internet in
the United Kingdom. A paper presented at the GIS Research UK 6th National Confer-
ence, 31 st March - 2nd April 1998, Edinburgh UK.
Smith, T.E. 1989. Shortest path distances: An axiomatic approach. Geographical Analysis,
21:1-31.
Stewart, J.Q., and Warntz, W. 1958. Physics ofpopulation distribution. Journal 0/ Regional
Science, 1:99-113.
Tobler, W.R. 1976. Spatial interaction patterns. Journal 0/ Environmental Systems, 6:271-
301.
Tobler, W.R., 1978. Migration fields. In Clark, W.A.V. and Moore, E.G. (eds.) Population
Mobility and Residential Change. Northwestern University Studies in Geography Num-
ber 25:215-32.
Tobler, W.R., and Wineburg, S. 1971. A Cappadocian speculation. Nature 231 :39-41.
Watts, D.J., and Strogatz, S.H. 1998. Collective dynamies of 'small world' networks. Na-
ture 393:440-42.
Webber, M.M. 1964. The urban place and the non-place urban realm, in Webber, M.M.,
Dyckman, lW., Foley, D.L., Guttenberg, A.z., Wheaton, W.L.c., and Wurster, C.B.
(eds.) Explorations into Urban Structure. Philadelphia PA: University of Pennsylvania
Press, 79-153.
9 Who's Up? GlobalInterpersonal Temporal
Accessibility
9.1 Introduction
A petroleum executive sitting at her desk in Houston knows when a satellite phone
call might find her chief geologist hard at work in Uzbekistan, but this is nothing
new. Since the end of the World War II, many a British veteran has maintained
scheduled weekly contact with harn radio comrades half-a-day away in Austral-
asia. And long before the Tokyo Stock Exchange operated on a 24-hour cycle,
brokers in New York City knew when to reach colleagues at the office in Japan; if
they weren 't certain, all they had to do was glance up at one of the half dozen or
so clocks mounted on the wall. Even new initiates to the World Wide Web learn
quickly that file transfers can be expedited by downloading from sites in countries
where most of the population lies dormant. Computer-mediated communication
across time zones surely must reach its zenith in emergency tele-medicine:
interactive specialists in Toronto, Paris, and Auckland confer while viewing
digital injuries sent from backpack transmitters by doctors on the ground in war-
tom Bosnia. Fine and weil, but when might a Japanese child log on in the
classroom to chat with a virtual pen-pal in Canada? Chances are, never. Even in
the near future of affordable webcams, speech recognition, and simultaneous
translation, it will be impossible for students in Halifax to converse with si ster-city
pupils in Hakodate. Why? When Japanese students start school at 9 o'clock in the
morning, it is 10 o'cIock at night on Canada's east coast. Simply put, one city or
the other is always going to be asleep. So obvious, perhaps, are such points that
they have largely escaped the gaze of academic researchers in the social sciences.
The ability of individuals to interact in real-time around the planet by way of the
Internet forms the basis ofthis paper. Beginning with the assumption that the time
of day will always be one of the few things truly separating the world's cultures,
we explore interpersonal accessibility with a particular emphasis on the role of
daily activity patterns. At this early stage in our investigation, we conceptualize a
'personal real-time accessibility index' to measure the potential for direct,
immediate, and responsive face-to-face global remote interaction. Following so me
discussion of research in accessibility and of studies in time use, a case study of
148 A.S. Harvey and P.A. Macnab
There has been increasing concern over an impending decline in social interaction.
Several factors continue to stimulate this concern. First is the declining role of the
workplace, at one time, the main nucleus for networking and social mobility in
modern society. Next is the escalating complexity related to the synchronization of
individuals' time budgets. This colonization oftime, as it has come to be known, is
expected to deter social interaction considerably. A future society might be envi-
sioned in which individual units remain connected through cyberspace yet isolated
in time and geographic space. This can be seen as a dangerous trend since social
interaction has been a strategy of individual survival both within and away from
the workplace. In spite of spatial separation, emerging communication technolo-
gies enhance the ability of individuals to share experiences and information. Com-
puters, for example, '... are being used more and more ... for all types of
communication, which traditionally have taken, place face to face' (Batty 1997).
The most timely and meaningful exchanges still require personal accessibility and
direct interaction in real-time. While phone systems and conference calls most
often enable these exchanges, various Internet tools such as 'instant messaging'
are gaining in popularity.l Significant baITiers do, however, still exist to such real-
time sharing of experiences and information at the global level.
As numerous authors note, the concept of accessibility is easy to use and hard to
define (Pirie 1979). We will not enter that debate in this paper except to indicate
that by accessibility we are interested primarily in access to people (e.g., telephon-
ing someone) and not in access to places (e.g., lunch at an exclusive country club).
Two particular aspects of accessibility that have been identified in the literature
are drawn upon for our explorations: first is the notion of connectivity, and second
the notion of potential (Fellman, Getis, and Getis 1990). The Internet provides
connectivity. What needs to be determined is the potential for individual interac-
tion.
Interpersonal temporal accessibility (IT A) or interpersonal reachability means
an individual is available to interact reciprocally with other individuals in the
normal course of daily activity. In its simplest form, the personal real-time acces-
1 In a discussion of the proliferation of IRC sites (Internet Relay Chat), Rheingold (1993)
characterizes the assembled interlocutors as real-time tribes.
GlobalInterpersonal Temporal Accessibility 149
sibility index is the population reach (i.e., the effective number of persons reach-
able) from a given location at any given time. Extensions could be the role group
(e.g., worker, student, other) and possibly the nature ofthe contact. The index var-
ies both spatially and temporally. The potential for IT A in so me locations greatly
exceeds the potential in other locations. For any given location, potential will vary
over the diurnal cycle as distant locations and peoples become more or less acces-
sible.
In its simplest form, the ITA index at any hour ofthe day can be given as:
24
IA="'a*u*P
L..J } } } (9.1)
j=1
Where:
Hence, the total potential population reach from a given time zone is the sum of
the number of persons available for interaction in all of the time zones. This
measure shows only how many individuals can be contacted with no indication of
the duration of time they are available for communication. The total time avail-
ability can be caJculated as the duration (d) oftime persons are alone and awake:
Personal Accessibility
Giddens (1990) identifies two types of social interaction: (1) face-to-face and,
(2) remote. Both transport and communications mediate the latter, identified as
time-space distanciation. Since the rapid diffusion of information technology has
considerably diminished the friction of distance as an impediment to interaction
(Graham and Marvin 1996, Cairncross 1995), the present investigation will focus
on the role played by communication. In Janelle's (1995) view, communication
can require temporal coincidence, or not, and spatial coincidence, or not. Various
possibilities are shown in Iable 9.1.
parties
No Refrigerator notes C Answering and D
Hospital charts recording machines
required MaillE-mail
Telegrams, telex, fax
Printed publications
Computer conferencing
and D). Real-time distanciation indicates that the means of communication en-
ables an immediate exchange of information. Such an exchange might be thought
of as mediated face-to-face contact. 2 Delayed distanciation involves aseparation
in both time and space. Wehave three meaningful categories with respect to
communication:
The remainder of this investigation is concerned with the second category, me-
diated face-to-face interaction. Our focus is communication, rather than transpor-
tation, and temporal constraints, rather than spatial constraints. Temporal
constraints take two forms and the significance of each must be recognized. First,
temporal location determines when activities occur (e.g., travel to work in the
morning). Second, duration determines the amount of time needed (e.g., viewing a
movie requires about two hours). In order for mediated face-to-face contact to take
place, individuals must be free of both constraints; that is, they must be temporally
accessible at the required moment and have sufficient time available to interact
meaningfully with others.
Hägerstrand (1970) posits three sets of constraints that will be drawn upon for our
discussion ofaccessibility: (1) capability, (2) coupling, and (3) authority. Capabil-
ity constraints are imposed by physiology, abilities and access to tools. Coupling
constraints define when, where, and for how long one needs to join with other
individuals. An individual or group imposes authority constraints through control
of a time-space entity. Each constraint plays a role in determining global ITA. The
increasing availability and affordability of computers are enhancing the capability
for global ITA by way of the Internet. Although baITiers have been greatly dimin-
ished, they are far from having been eliminated (Schuler 1996, Salomon 1988).
There is still a need for equipment and software (a vehicle) as weil as an access
point and time (an on-ramp andfuel).
2 In asense, this is analogous to an encounter between a deaf person and a hearing person
where a third party interprets with sign language to facilitate an exchange of information
in the present.
152 A.S. Harvey and P.A. Macnab
Capability Constraints
The ability to communicate randomly via the Internet demands personal and tech-
nological capabilities. Personal capabilities comprise adequate computer skills and
language abilities - possibly foreign languages - while technological capabilities
entail the availability of a computer, communications software, and some kind of
network interface (e.g., modem or Ethernet card). Table 9.2 illustrates these re-
quirements and others with a corresponding measure of constraint for three modes
of communication: text, voice, and full audio-visual. Successive levels of interac-
tion require more elaborate and costly devices (i.e., for voice a microphone is re-
quired, and for video, a camera).3 The degree of constraint varies by device and
mode; for example, the exchange oftext is minimally affected by computer power,
but moderately affected by the need for compatible communications software.
Capability constraints, measured by access to appropriate technology, represent
the initial constraint on accessibility. This is true for two reasons. First, the tech-
nology must be in place for communication to take place. Second, the develop-
ment of personal skills will follow the availability of technology since computing
skills are primarily learned through practice. Hence, the speed with which popula-
tions become potentially accessible depends on the three factors: (l) the speed of
diffusion of the technology, (2) the speed of adoption by individuals, and (3) the
length of the learning curve to bring one online in an active fashion. Understand-
ing each of these factors is in itself a study that we will leave for future explora-
tions.
Language differences remain a significant barrier to real-time interpersonal
communication. Individuals must share a common spoken or written form of lan-
guage if they are to take full advantage of existing communication technologies.
The barrier posed by language will diminish as voice recognition and translation
software improves, but for the present discussion, language abilities are taken to
be aprerequisite for functional communication. 4 Even without considering the
built-in bias towards Roman characters, the dominance ofEnglish language on the
Internet, trailed at some distance by Japanese, German, and Spanish (see Table
9.3), must be regarded as a major constraint to widespread communication. While
there is a strong possibility that the numbers presented in Table 9.3 fail to account
for the use of English by non-native speakers, there can be little doubt that the use
of English by non-anglophones far exceeds the use of foreign languages by native
speakers of English.
3 Combinations of levels exist as weH, e.g., in the burgeoning online adult entertainment
trade, aperformer often responds to customers' typed requests with a one-way video feed.
4 Consider the manner in which various technologies have reduced barriers for persons with
hearing, visual and motor impairments. For example, optical character recognition en-
ables scanned text to be converted and enlarged on the screen or embossed in Braille;
voice generation software reads from computer files; and voice recognition tools capture
lectures in print and word prediction prograrns to reduce keyboarding.
GlobalInterpersonal Temporal Accessibility 153
Coupling Constraints
5 Portable phone systems in all of their forms continue to have an undeniable effect on
when and where people can be reached.
GlobalInterpersonal Temporal Accessibility 155
Ideally, one would divide social circles into at least three categories: (1) family,
(2) friends, and (3) other acquaintances. However, existing data and survey limita-
tions require that the social circles be generalized down to a pair. First is the fam-
ily circle, the principal unit of social interaction in society. This circle includes
spouse or partner, children, and any other family member, whether they live in the
household or not. Second is the circle of other acquaintances; any person with
whom individuals interact who is not related by blood or marriage.
Social space, physical surroundings or location can be meaningfully identified
by three main venues. First is horne, the household or place of residence. The sec-
ond encompasses work or education sites, professional associations, and office or
school clubs. Third is community space, composed of social or recreational clubs
(not related to work or schooI), church, community organizations, restaurants,
cafes, and houses of friends and relatives.
Preliminary research has shown that the segregation of social circles is not eas-
ily accomplished. In reality they intermingle, at times becoming the single entity
generally referred to as society (Schneider, 1972; Harvey et al. 1997). Still, with
time-use data it is possible to formulate an operational measure of social contact
for analysis purposes. Three types of social interaction are integrated with social
space in Table 9.5 to create a social environment matrix (Harvey et al. 1997).
Since an individual's availability is a function of both his or her social space and
social circle, the social environment provides a framework for evaluating the
availability of individuals.
In our view time spent awake and alone is the significant factor in individual
accessibility. As Hägerstrand argues,
when an activity has started or a room has become occupied these cells become closed
for some duration of time ... projects looking for the same locations in space and time
have to await their turn or go elsewhere (Hägerstrand 1973, 80).
Individuals will occasionally interrupt conversations or activities to interact with
others, but this is generally viewed as unacceptable; therefore, a primary require-
ment of IT A is being alone. Location must also be considered since the mere act
of being alone will not necessarily guarantee accessibility. Assuming contact over
the Internet, it is reasonable to assurne further that individuals will normally have
access primarily at horne or work. We readily acknowledge airplane modems,
portable e-mail addressing, Internet cafes, public terminals, and community-
oriented networks, but for the present discussion, we shall regard such forms of
156 A.S. Harvey and P.A. Macnab
Authority Constraints
Authority constraints are less obvious and less amenable to quantification than
those re1ated to capability and coupling. Employer's banning of Internet usage for
personal communication during working hours is a growing example of this type
of authority. Other constraints emanate from religious prohibition and ritual ob-
servance, both of which usurp time available for accessing others. The V -chip,
encryption technology and World Wide Web content filters have all been devel-
oped to limit access to sites and material by specific groups. This rapidly evolving
area requires further consideration and analysis in the context of IT A. 7
6 In a survey conducted more than two years ago, Schuler (1996) uncovered nearly 300
computer systems designed exclusively for community use and hundreds more in the
planning stages. While most community networks offer Internet access through commu-
nity-based computing centers, many are also interested in rearguard technologies such as
universal voice-mail, local news provision, community-based radio stations, and local ca-
ble television. Several federal governments, including Canada's, continue to embrace cuf-
fing edge communication technologies before less advanced forms reach their potential
(e.g., see Information Highway Advisory CounciI1995).
7 For an extended treatment of authority-related issues in computerization, see the collec-
ti on edited by Kling (1996).
Global Interpersonal Temporal Accessibility 157
Capability Measurement
35 % ]
30 %
~ 25 %
S
. Home
~ 20 % i i
~ _ Work
~ 15 % . _O ther I
~
~ 10 %
5%
0%
,$<:-
q.,if' ,y'"
~
Time-use Data
Time-use studies provide ample evidence of temporal variation in the city's social
geography due to activity patterns and settings (Goodchild and Janelle 1984, Har-
vey et al. 1997, Janelle, Klinkenberg, and Goodchild 1998). While there is an ex-
treme dearth of time-space data, time-diary data do exist for a large and growing
number of countries. Studies typically capture extensive information about what
each participant is doing, where they are, and who they are with. People ac count
for every minute throughout a given recording period, usually 24 or 48 hours.
While most of the existing databases lack detailed geographical coordinates, indi-
viduals can usually be identified by region, state, or province. Time zones, the
geographical characteristic of most relevance for the present undertaking, can be
determined readily from these fields.
Most time-use studies provide some spatial detail of activity and generic loca-
ti on (e.g., horne, workplace, other place, or traveling). Generic locations are a ma-
jor indicator for the type of contact likely to be made with respect to connections
at horne, at work, or in the community. Data drawn from time-use studies, con-
ducted by central statistical offices in Canada, Norway, and Sweden, provide
numbers describing involvement in various social environments. People spend
approximately 950 to 1,000 minutes of the day awake (see Table 9.7). Slightly
over 500 minutes of this time is spent in the household and between 170 and 250
minutes are spent alone in the workplace. Ofthe remaining time, 146 to 186 min-
utes is allotted to activities in the community, while 70 to 80 minutes are spent in
transit.
The extent to which IT A is affected by the diurnal cyc1e is amply illustrated with a
case study of Canada, one of the few countries to span six time zones. The re-
mai nd er of this investigation draws primarily on time-use data collected by Statis-
tics Canada in 1992 during Cyc1e 7 of the General Social Survey Program. Figure
9.2 depicts the provincial boundaries and time zones ofthe country.
Provinces
and
Time Zones
of
Canada
,i ' CO r:"
'. ..
Sl John"s
0?
u (f?
Pacific
Mountain
~
Centra l
Eastern Atlantic
100
90
80
70
Soclal Environment
60
~ • (xh«• • (X1"!er
c:
0 . . (Xh« • •Work
50
,---
p
ro
"5
Cl.
0 40
.!.! .! (Xhef • • Homo
n.
AIone.T'.....
JO
20
10 AIone,Home
Timezones
Figure 9.3. Social environment across Canada at 10:00 a.m. Newfoundland time.
8 Liberty has been taken in the example by treating Newfoundland as if there was a full
hour difference from the Atlantic time zone when the difference is only a half-hour.
9 The totals depicted here are raw percentages. In actual numbers of people, 50 percent of
the eight million Ontarians represents a far greater population than 95 percent of the one
million Newfoundlanders.
GlobalInterpersonal Temporal Accessibility 161
100
00
80
70
Soclal Environment
60 Others, Other
c:
0
~ so
'3
0-
0
a.
40
30
20
A.Ione ,\IVofk
10 AJone,Home
0
Nfld AlIMtIc Eastern Cenlrlll Moo.niW> P""fic
TImezones
Figure 9.4. Social environment across Canada at 3 :00 p.m. Newfoundland time.
trafik monitoring (e.g., Batty 1997, Jiang and Ormeling 1997, Dodge 1996) are
two areas in which the mapping oftime has been applied extensively.
--
160000u
~ 120000u
~ - r---
~ 80000n. r---
0
"iii
III - r---
c
0 r---
~ 40000LT
&. - '---
•
Time of day in Newfoundland
Other Work Home
D
Figure 9.5. Canadian ITA adjusted to Internet usage showing Internet-accessible Canadians
at selected times far Newfoundland.
Spatial changes over time are weil represented on maps, but what about time
changes over space? How often has a phone book been opened to check interna-
tional time zones on the long-distance pages?IO Classic representations, typified by
static maps and textual descriptors, have been augmented in recent years by a
range of software and online conversion tools (e.g., CLOX World Time Zone
10 Say I wish to call a cousin in New Zealand and the phone book teils me that her country
is Atlantic Standard Time plus 14 hours. That means ifit's 8 o'clock in the evening here,
it's 10 o'clock in the morning there -- dam, she's already left for work!
Globallnterpersonal Temporal Accessibility 163
Table 9.8. Sodal environment across Canada, 3:00 p.m. Newfoundland time
TIMEZONES
Nfld. Atlantic Eastern Central Mountain Pacific Total
3:00 2:00 1:00 12:00 11:00 10:00
Sieeping Count 10218 43656 297772 19788 107016 130320 608770
Row% 1.7 7.2 48.9 3.3 17.6 21.4 100.0
Col% 2.3 3.2 2.2 2.4 4.0 5.0 2.9
% Total .0 .2 1.4 .1 .5 .6 2.9
Alone, Count 95865 329778 2989528 151224 584226 651540 4802161
Horne Row% 2.0 6.9 62.3 3.1 12.2 13.6 100.0
Col% 21.7 23.9 22.4 18.0 21.8 25.1 22.6
% Total .5 1.5 14.0 .7 2.7 3.1 22.6
Alone, Count 10868 50770 668205 17781 114694 125613 987931
Work Row% 1.1 5.1 67.6 1.8 11.6 12.7 100.0
Col% 2.5 3.7 5.0 2.1 4.3 4.8 4.6
% Total .1 .2 3.1 .1 .5 .6 4.6
Alone, Count 19902 56558 733251 27372 96565 124243 1057891
Comm. Row% 1.9 5.3 69.3 2.6 9.1 11.7 100.0
Col% 4.5 4.1 5.5 3.3 3.6 4.8 5.0
% Total .1 .3 3.4 .1 .5 .6 5.0
Alone, Count 18133 55526 692708 43836 94853 94766 999822
Transit Row% 1.8 5.6 69.3 4.4 9.5 9.5 100.0
Col % 4.1 4.0 5.2 5.2 3.5 3.6 4.7
% Total .1 .3 3.3 .2 .4 .4 4.7
Others, Count 88539 204690 1974478 165648 370461 295884 3099700
Horne Row% 2.9 6.6 63.7 5.3 12.0 9.5 100,('"
Col% 20.1 14.9 14.8 19.7 13.8 11.4 14.6%
% Total .4 1.0 9.3 .8 1.7 1.4 14.6%
Others, Count 86194 293240 2813188 203551 729583 660674 4786430
Work Row% 1.8 6.1 58.8 4.3 15.2 13.8 100.0
Col% 19.5 21.3 21.1 24.2 27.3 25.4 22.5
% Total .4 1.4 13.2 1.0 3.4 3.1 22.5
Üthers, Count 111550 343366 3193094 210922 578856 513819 4951607
Other Row% 2.3 6.9 64.5 4.3 11.7 10.4 100.0
Col% 25.3 24.9 23.9 25.1 21.6 19.8 23.3
% Total .5 1.6 15.0 1.0 2.7 2.4 23.3
Total Count 441269 1377584 13362224 840122 26762542596859 21294312
Row% 2.1 6.5 62.8 3.9 12.6 12.2 100.0
Col% 100 100 100 100 100 100 100
% Total 2.1 6.5 62.8 3.9 12.6 12.2 100
Source: Calculatedfrom Statistics Canada, General Social Survey, Cycle 7.
164 A.S. Harvey and P.A. Macnab
Clock, ID LOGIC World Time Zone Clock for Windows). Interactive time zone
converters vary in their utility and mode of global visualization, but generally,
most ofthe packages appear to permit queries such as: ifit is 6:00 p.m. in country
A, what time is it in country B? Some converters are text-based while others per-
mit graphie database queries, a useful function likely inspired by GIS. l1 Figures
9.6 and 9.7 offer preliminary cartographic depictions of IT A in Canada at 8:30
a.m. and 7:30 p.m. Newfoundland time.
11 At least one software package provides a visual indication of who' s awake. The Moon
and Earth Viewer offers a map or image of the world shaded to indicate which countries
are experiencing daylight and which fall under the shadow ofthe moon.
GlobalInterpersonal Temporal Accessibility 165
SO~I Environment (}
Number of People
Alone and Awake
__ Horne
..._-0::::--- 914 ,278
~
. . Other
. . Work a 8~ I:b Q
D~~
_______---;'-- 107 ,208
,.-..::-+--7'-- 25 ,700
Pacific
Mountain
Central Atlantic
Eastem
Figure 9.6. Preliminary cartographic depiction ofITA in Canada, 8:30 a.m., Newfoundland
time
166 A.S. Harvey and P.A. Macnab
SO: I Environment
_ Horne
. . Other
. . Work
~
aa~ ~Q
Q
Accessibility at 7:30pm ewfoundland Time
Number of People
Alone and Awake
.,_-0:::--- 529,311
------+--
D3~
190 ,9 14
........,---IL+- 28,569
Paci fic
Atlantic
Mountain
Central
Eastern
Figure 9.7. Preliminary cartographic depiction ofITA in Canada, 7:30 p.m., Newfoundland
time
Our preliminary inquiries suggest research needs in five principal areas: (1) inter-
active geographie visualizations, (2) diminishing accessibility as a result of in-
creasing accessibility, (3) Internet-traffic monitoring to compare IT A potential
with real-time communication volumes, (4) expansion of time-use and Internet
surveys to collect detail on the nature and duration of mediated face-to-face inter-
action, and (5) continued exploration around the notion oftemporal regions within
and between the planet's geographie regions.
would show Canada waking up. The potential for further visualization in-
c1udes several interactive GIS add-ons that would enhance the options de-
scribed above for time-zone converters. What if you wished to determine
where and when the most people might be reached? By extending a time-zone
database to inc1ude diurnal activity patterns at a larger scale (e.g., telephone
area codes or Internet-service providers), one could query the statistically op-
timal time of day for accessibility between two cities. Conversely, one could
determine the location where the maximum number of people could be
reached at a given time. The World Time Zones data set available from ESRI
contains 35 zones at a scale of 1:3 million. Coded to indicate the hours of dif-
ference from Greenwich Mean Time, this coverage might be a useful starting
point. We anticipate and welcome any additional suggestions for GIS treat-
ment of these connections across time and space.
(3) The exploratory calculations presented here offer one possible index of inter-
personal-temporal accessibility. What other measures should be factored in?
The numbers derived for potential accessibility on the Internet across Canada
bear comparison with actual counts and duration volumes of mediated real-
time connections. At present, it is difficult to extract these measures from
electronic-traffic reports; but, in time, such monitoring is expected to become
commonplace. Are the numbers embedded in time-use datasets useful and ac-
curate for predicting accessibility? If so, might call centers, community
groups, fundraisers, NGOs, entertainers, retailers, educators, business inter-
ests, and media outlets around the world be keen to make use of the numbers?
9.7 Summary
Acknowledgements
Staff ofthe Time Use Research Center at Saint Mary's University assisted at several stages
during the course of our investigation. We are grateful for the contributions of Jennifer,
Barbara and Wendy.
References
Allardt, E. 1990. Challenges for comparative social research. Acta Sociologica 33(3):183.
Barker, R.G. 1951. One Boy 's Day: A Specimen Record of Behaviour. New York: Harper.
Batty, M. 1996. Visualizing Urban Dynamics. In Longley, P. and Batty, M. (eds.) Spatial
Modelling in a GIS Environment. Cambridge UK: Geoinfonnation International, 297-
320.
Batty, M. 1997. Virtual geography. Future 29(4/5):337-52.
Cairncross, F. 1995. The death of distance: a survey oftelecommunications. The Economist
336(7934):5-28.
Dodge, M. 1996. Mapping the World Wide Web. GIS Europe 5(9):22-4.
Egenhofer, M.J. and Golledge, R.G. (eds.). 1998. Spatial and Temporal Reasoning in Geo-
graphic Information Systems. New York: Oxford University Press.
Fellmann, J., Getis, A and Getis J. 1990. Human Geography, Second Edition. Dubuque IA:
Wm. C. Brown.
Giddens, A. 1990. The Consequences ofModernity. California: Stanford University Press.
Goodchild, M.F. and Janelle, D.G. 1984. The city around the clock: space-time patterns of
urban ecological structure. Environment and Planning A 16:807-20.
Graham, S. and Marvin, S. 1996. Telecommunications and the City: Electronic Spaces,
Urban Places. London: Routledge.
Hägerstrand, T. 1970. What about people in regional science? Papers and Proceedings of
the Regional Science Association, 24:7-24.
Hägerstrand, T. 1973. The domain ofhuman geography. In Chorley, R.J. (ed.) Directions in
Geography. London: ethuen & Co. Ud, 67-87.
Harvey, AS. 1997. From activities to activity settings. In Ettema, D. and Timmennans, H.
(eds.) Activity-based Approaches to Travel Analysis. Tarrytown NY: Pergamon.
Harvey, AS. 1982. Role and context: Shapers ofbehaviour. Studies ofBroadcasting 18:70-
92.
Harvey, AS., Taylor, M.E., Ellis, S. and Aas, D. 1997. 24-Hour Society and Passenger
Travel. Report cmmissioned by Transport Research Centre, Ministry of Transport, Public
Works and Water Management. Rotterdam.
Infonnation Highway Advisory Council. 1995. Access, Affordability and Universal Service
on the Canadian Information Highway. Ottawa: Supply and Services Canada.
Janelle, D.G. 1995. Metropolitan Expansion, Telecommuting, and Transportation. In S.
Hanson, S. (ed.) The Geography of Urban Transportation, Second Edition. New York:
Guilford Press, 407-34.
Janelle, D.G., Klinkenberg, B. and Goodchild, M.F. 1998. The temporal ordering ofurban
space and daily activity patterns for population role groups. Geographical Systems 5: 117-
37.
Jiang, B. and Onneling, F.J. 1997. Cybennap: The map for cyberspace. The Cartographic
Journal 34(2): 111-16.
170 AS. Harvey and P.A Macnab
Kling, R. (ed.). 1996. Computerization and Controversy: Value Conjliets and Social
Choiees. Second Edition. San Diego: Academic Press.
Langran, G. 1992. Time in Geographie Information Systems. Bristol PA: Taylor and Fran-
cis.
Lewin, K. 1951. Field Theory in Social Scienee: Seleeted Theoretieal Papers. New York:
Harper.
Neilsen Canada Ud. Personal communication, 1998.
MacEachren, AM. 1995. How Maps Work: Representation. Visualization and Design. New
York: The Guilford Press.
Maslow, AH. 1954. Motivation and personality. Harper's Psyehologieal Series. New
York: Harper.
Pirie, G.H. 1979. Measuring accessibility: A review and proposal. Environment and Plan-
ning A:ll:299-312.
Rheingold, H. 1993. The Virtual Community: Homesteading on the Eleetronie Frontier.
Reading MA: Addison-Wesley.
Salomon, I. 1988. Geographical variations in telecommunications systems: The implica-
tions for location of activities. Transportation 14:311-27.
Schneider, A 1972. Patterns 0/ Social Interaetion. The Use 0/ Time: Daily Aetivities 0/
Urban and Suburban Populations in Twelve Countries. The Netherlands: Mouton & Co.
Schuler, D. 1996. New Community Networks: Wired/or Change. Reading MA: Addison-
Wesley.
Statistics Canada, General Social Survey, Cycle 7, 1992.
Vasiliev, I. 1996. Design issues to be considered when mapping time. In Wood, C.H. and
Keller, c.P. (eds.) Cartographie Design: Theoretieal and Praetieal Perspeetives. New
York: John Wiley and Sons, 137-46.
10 The Role of the Real City in Cyberspace:
U nderstanding Regional Variations in
Internet Accessibility
10.1 Introduction
Since 1993, when the first graphical web browser, Mosaic, was released into the
public domain, the Internet has evolved from an obscure academic and military
research network into an international agglomeration of public and private, local
and global telecommunications systems. Much of the academic and popular litera-
ture has emphasized the distance-shrinking implications and placelessness inher-
ent in these rapidly developing networks. However, the relationship between the
physical and political geography of cities and regions and the virtual (or logical)
geography ofthe Internet lacks a strong body of empirical evidence upon which to
base such speculation.
This chapter presents the results of aseries of studies conducted from June 1996
to August 1998. Our research suggests there is a metropolitan dominance of Inter-
net development by a handful of cities and regions. We identifY and describe an
emerging structure of 'virtual' hubs and pathways which are linking a set of major
cities in the United States, suggesting that there is a complex emerging inter-urban
communications network that goes far beyond Castells' (1989) informational
mode of development.
More importantly, we analyze the utility and relevance ofthree measurements of
Internet development for urban planning and policy research. The first measure is
the number of computers connected to the Internet on a full-time basis, based on
data collected by Matrix Information and Demography Services of Austin, Texas.
We find this measure to be of limited usefulness, as it only measures location-
specific hardware installations with little reference to their purpose or function.
The second measurement system we explore is based on the Internet's addressing
scheme, known as the domain name system. Unique to individual organizations
(business, education, non-profit, government), domain names are registered with
InterNIC, an administrative clearinghouse contracted by the National Science
Foundation. Associated with each domain name is a unique billing address, which
172 M.L. Moss and A.M. Townsend
permits the localization of the organization using that name. This is the most in-
formative measure for assessing variations in Internet use across a variety of geo-
graphie units, from states to individual ZIP codes. Finally, we examine the
capacity and topography of nationwide Internet backbone networks that transport
data between metropolitan areas.
Based on these three measurements, we find that a limited number of cities and
metropolitan areas dominate the rapidly emerging telecommunications landscape
of the United States, leading in the development of increasingly sophisticated ap-
plications and technologies. Accessibility to the most highly developed real and
virtual Internet infrastructure is a metropolitan phenomenon, and highly stratified
among regions and cities. Furthermore, we describe each of the three measure-
ments used - host counts, domain counts, and backbone network capacity - and
their unique advantages and disadvantages for research. Successfully applying
them to urban analysis requires not only an understanding of urban and regional
development processes, but also the purpose, design, and ftmction of these com-
plex technical systems.
The primary question this series of studies seeks to address is which cities and
metropolitan areas exhibit a rapid buildup of Internet-related telecommunications
infrastructure, and secondly, how can we effectively measure the geographie dis-
tribution of the Internet? Measuring this phenomenon proved very difficult. First,
as noted earlier, telecommunications providers have strong incentives to keep data
on their operations closely guarded. Second, from a very early stage in its devel-
opment the Internet was designed to subvert geography by using a packet-
switched message routing system, which operates more like the postal service than
the telephone network. As a result, messages can be rerouted around faulty
switches or stations. Thus, the few aspects of the Internet that can be localized and
measured have only limited relevance to physical geography. Third, each of the
three data sources measured aspects of the Internet that did not always have over-
lapping geography. Finally, we needed to consider what these measurable criteria
could indicate about the level of human activity on the Internet in different cities
and regions and the socioeconomic consequences of these information flows. This
section describes the three measurements we used and the strengths and limita-
tions of each. The final section presents a composite picture of the North Ameri-
can system of cities we have derived from these observations.
Matrix Information and Demography Services of Austin, Texas has measured the
geography of the Internet since the early 1990's, and is considered a leader in ana-
Iyzing what it calls the matrix of inter-connected global computer networks.
MIDS' methods permit the localization oflnternet hosts (computers connected to
the Internet on a full-time basis) to geographie areas as specific as astreet address.
Our study was based on data provided by MIDS for January 1996, and indicated
the number of Internet hosts per county for all 50 states. The data included a sub-
stantial number of manual corrections to the automated survey that generates the
data, based on MIDS' proprietary knowledge of the known geographie location of
large clusters of host computers in corporate research and development centers,
for example.
This measurement was a useful first step towards visualizing the emerging tele-
communications landscape of the United States. As anticipated, Santa Clara
County, California (Silicon Valley) and Middlesex County, Massachusetts (Route
128) were the largest clusters of permanently connected Internet hosts. Table 1
shows the top 25 counties by number of host computers. However, we did not
anticipate the large number of central cities that appear on this list. Furthermore,
12 of the 25 top counties were located within just four metropolitan areas: San
Francisco, Los Angeles, Washington DC, and New York. An asterisk indicates
these twelve counties. While these results confirmed our suspicions regarding the
The Role ofthe Real City in Cyberspace 175
At this point, we encountered the first of many instances in which the techno-
logical realities of the Internet drastically affected the interpretation of our results.
The measurement of host counts was extremely coarse because it failed to differ-
entiate between various types of computer equipment. For example, if a financial
services company in Manhattan uses a highly centralized computer system based
on a mainframe and durnb terminals, this method might only count a single host -
the mainframe being the only machine direcdy connected to the Internet. Con-
versely, a small software company in Silicon Valley that uses a Local Area Net-
work to connect its dozen microcomputers to the Internet individually would have
a much higher host count. In some cases, networked printers might even be
counted by this method. Finally, the increased use of firewalls, computers de-
signed to mediate extern al Internet connections and shield institutional networks
from intruders, excludes a significant number of hosts from detection, especially
those of large corporations. These factors generate significant variations in the
effectiveness of comparing host counts across regions and industrial sectors.
It also seemed counter-intuitive to use a measurement based purelyon the tech-
nical organization of the Internet to infer some understanding of the rate of adop-
tion of these technologies across regions. The nature of the modem American
economy and the production systems of the Internet services industry further
complicates this approach. Many organizations do not physically house their
Internet-accessible information at their physical location, preferring to hire con-
tractors who provide expertise and equipment. While the information-producing
jobs and economic activity associated with a website may take place at a central-
ized office in a dense urban area, it is just as likely that the fruits of this labor are
electronically disseminated from a remote location, which could conceivably be
located anywhere. These limitations in the host count measurement led us to seek
other indicators of Internet use, which proved more useful in understanding the
spread ofthe Internet among cities.
Domain names are one of the basic forms of Internet addressing, which map
groups of numeric Internet addresses to intuitive names like nyu.edu or alt.com.
Each domain name is registered with Network Solutions, an organization char-
tered by the National Science Foundation to administer the domain name system.
As Figure 10.1 shows, each name is registered to an individual or organization,
and the publicly available registration re cord contains a billing address for that
entity. From this information, it is possible to localize the location of the entity
that owns that domain name as specifically as the postal code (ZIP) level. This
geographical specificity of the domain name makes it a highly attractive measure
for Internet activity. Network Solutions has enjoyed a monopoly over domain reg-
istrations for the most popular commercial, non-profit, educational, and govern-
ment domains in the United States, leaving only a small portion of the American
Internet beyond the scope of these data.
The Role ofthe Real City in Cyberspace 177
The strengths of this measurement for urban research stern from its representa-
tion of a social phenomenon, rather than a teehnieal one. Beeause eaeh domain
name roughly eorresponds to a corporate, government, or edueational entity, this
measurement indieates spatial variations in the adoption and use of Internet-based
eommunieations by organizations. Sinee nearly 90 percent of Internet growth over
the period between 1994 and 1997 was from the addition of eommereial domain
names, these results primarily measure the extent to whieh businesses deployed
these new teehnologies. Domain registrations also indieate the date eaeh domain
was first registered, permitting us to identity those regions that had the most rapid
growth in Internet use.
Registrant:
Brooklyn, NY 11211
US
those of the company's products, or variations upon the company's name. This
practice may be responsible for some distortion in the overall results.
The primary source of data for our research on the geographic distribution of
domain names was Imperative! of Pittsburgh, Pennsylvania. Imperative! sells
mailing lists to direct marketers who offer Internet-related products to the owners
of domain names and maintains a database of currently registered domain names
that can be aggregated by nearly any geographical unit. We analyzed regional
variations in domain name registration at three geographie levels; (I) U.S. Census
regions, (2) 85 major U.S. cities, and (3) postal code areas in New York City
(Moss and Townsend I 997a, 1997b, 1998). Among major U.S. cities, the largest
clusters (Manhattan and San Francisco), were also the most densely networked.
Furthermore, in the 15 cities with the largest number of Internet domains, which
accounted for 12.6 percent of all U.S. domain registrations in April 1994, new
domains were registered faster than other areas. By 1997, these 15 cities ac-
counted for nearly one-fifth (19.7 %) of all U. S. domain registrations. Clearly,
Internet technologies were being more rapidly deployed in major urban areas.
Outside the den se urban areas of the Atlantic, Pacific, and Gulf coasts, only Phoe-
nix and Chicago had a significant number of domains.
We also computed the density of domains with respect to population, or do-
mains per 1000 persons. Domain density was typically highest among cities whose
primary function was as aresort, government, or education center. Nodal cities
also showed high concentrations (Moss and Townsend 1998).
Comparing these results to a limited set of domain counts for approximately a
dozen cities in April 1994 permitted us to track the rate of domain growth over a
3-year period. These cities registered domains far faster than the national average,
adjusted for population. The growth rate of domains between 1994 and 1997 was
linked to a city's relative position in the national urban system, with the advanced
service centers growing most rapidly. This indicates a strong relationship between
information-intensive economic activity and early adoption of the Internet among
businesses. The cities primarily fall into four broad categories of growth rates,
summarized in Table 10.2.
The fact that the set of cities most commonly referred to as world eities or
global eities recorded the slowest growth rates among large Internet clusters is
disturbing, for it indicates an averaging function. As our detailed analysis of do-
main registrations in Manhattan indicates, adoption of the Internet is not a wide-
spread phenomenon across urban populations. Rather, it is almost entirely limited
to the central business districts, with moderate adoption rates in the more success-
ful immigrant communities (Moss and Townsend 1997a). The fact that smaller
cities such as Austin or Boston exhibit higher growth rates is most likely due to a
more even spread of technological opportunities among their more homogeneous
populations. On the other hand, world cities appear to be characterized by a digital
elite co-existing with a vast, largely disconnected information ghetto. As one ex-
ample of this disturbing trend, a preliminary survey of domain registrations in the
Los Angeles area indicates a much slower diffusion of Internet technologies
among Spanish-speaking and immigrant communities.
Jed Kolko, a doctoral student at Harvard University, is currently conducting
research that will extend the analysis of domain name registrations to all 285 U. S.
Metropolitan Statistical Areas (MSAs) over a four-year period from 1995-1998.
This work addresses a shortcoming of our research, in that we neglected to explore
Internet adoption and use in suburb an areas surrounding the major cities. In fact,
the bulk of new office space in recent decades has emerged not in the central cities
that we focused upon, but rather in the edge eities that surround them. (Garreau
1991) Some preliminary results, in Table 10.3, show that even where including the
surrounding metropolitan areas there is strong evidence of a select group of cities
that dominate Internet activity.
Once again, the four metropolitan areas that contained the largest clusters of
Internet hosts, and the densest concentrations of domains, also account for the
largest nodes ofInternet activity by this measure.
While it was important to identify centers of Internet activity and variations in the
concentration of Internet indicators among American cities, we also need to quan-
tify the flows of information between cities to understand how these networks are
developing within the American urban system. Communications on the Internet is
primarily carried over fiber optic networks, portions of which have been adapted
from their original use (the transmission of voice telephone caIls), although the
first dedicated networks optimized for the TCP/IP protocol are now under con-
struction. Traversing the country along traditional rights-of way, such as railroad
tracks, interstate highways, and even abandoned canals, these networks are the
physical manifestation of the 'information superhighway' . Like the host and do-
main-name measurement, our analysis of the conglomeration of national data net-
works collectively known as the Internet 'backbone' indicates a high degree of
centralization in the deployment and use of Internet technologies.
Again, the technology in question dictated careful interpretation of the results.
While the host count described a physical, real phenomenon (the connection of
computers to the Internet), and the domain count a thoroughly virtual, logical one
(the organization of Internet Protocol addresses into convenient hierarchies), the
measurement of backbone capacity is a hybrid. Although some networks operate
on isolated fiber optic cables, many are merely virtual networks operated over
lines leased from national and regional telephone companies. Often, a backbone
provider' s only capital equipment are the powerful routing computers that manage
the flow of data packets at network junctions (Rickard 1997).
While the geography and topography of these networks has received more atten-
tion from scholars than the identification of nodes of Internet activity, no other
studies have analyzed the aggregate topology and capacity of the major backbone
networks (Moss and Townsend 1998). Using maps and data from Boardwatch
Magazine's Quarterly Directory 01 Internet Service Providers, we compiled a list
ofthe capacity and endpoints of every major backbone link for 29 major backbone
operators in the United States. We estimate that these 29 wholesale providers sup-
ply at least 95 percent of long-haul Internet data transport services in the United
States. Based on the assumption that barriers to backbone accessibility were likely
to be found at the inter-metropolitan, rather than intra-metropolitan level, we fur-
ther aggregated these links by metropolitan area. The existence of networks, such
as Metropolitan Fiber System's Metropolitan Area Ethernets, and the rapid prolif-
eration of the baby BeIls' high-speed regional fiber networks underlies this as-
sumption. This aggregation allowed us to focus on the largest capacity fiber optic
networks, constructed of DS-3 (45 MBps), OC-3 (155 MBps) and OC-12 (622
The Role ofthe Real City in Cyberspace 181
MBps) technology. One megabit per second (MBps) indicates a data transfer rate
equal to approximately 128 pages of text per second.
The analysis is also confined to direct network connections. Theoretically, any
city on a given network has access to all locations served by all networks. How-
ever, there is significant traffic congestion of data packets at inter-network gate-
ways (bridges), and there are strong indications that providers have established
direct links on the most highly trafficked inter-metropolitan routes. For example,
several providers have established direct links between New York and Was hing-
ton, D.C., even though they already operate a route connecting these two metro-
politan areas through intermediate cities, such as Baltimore, Philadelphia or
Wilmington, Delaware. The San Francisco Bay Area is direct1y linked to almost
every metropolitan area in the United States, although many could presumably
have been served indirectly through another node. These patterns of investment
indicate the superiority of direct connections and their importance to a metropoli-
tan area's ability ofto import and export information via the Internet.
The results of the backbone analysis were the most striking and conclusive of
the three measurements we used. Summarized in Table 10.4, the data show that a
group of seven metropolitan areas (San Francisco/Silicon Valley, Washington,
D.C., Chicago, New York, Dallas, Los Angeles, and Atlanta) form a core group of
urban areas that dominate the Internet in the United States.
These seven metropolitan areas each have the capacity for over 5,000 megabits
per second (MBps) of Internet data throughput, sufficient to transfer text across
the Internet at a rate of over 640,000 pages per second. No other metropolitan ar-
182 M.L. Moss and A.M. Townsend
eas approach this level of capacity. Denver, ranked eighth, has only 60 percent of
the backbone capacity of seventh-ranked Atlanta. As a result, these seven metro-
politan areas share 62.0 percent of the nation's backbone capacity. The next 14
metropolitan areas together account for an additional 25.5 percent of the nation's
backbone capacity, while the remainder of the United States houses the remaining
12.5 percent. Outside the major metropolitan backbone hubs, communities are
linked to the Internet backbone via less robust data pipelines such as DS-1 lines
(1.5 MBps, also known as T-1), frame relay (0.05 to 0.25 MBps), ISDN (0.125
Mbps), and modem (0.025 to 0.5 MBps) lines, which substantially limits their
ability to move large amounts of information quickly. Many of these technologies,
especially the popular DS-1 lines, have notoriously distance-sensitive price struc-
tures that have severely restricted their proliferation in non-metropolitan areas.
These seven metropolitan Internet hubs also connect directly to a great variety
of other metropolitan areas and cities. The San Francisco Bay Area has the largest
number of external connections, with 153 links to dozens of other metropolitan
areas. Washington, Chicago, and New York have over 100 links maintained by
various network companies to other cities. Dallas, Los Angeles and Atlanta each
have 75 or more direct externailinks. By contrast, the next 14 metropolitan areas
each have 42 or fewer external linkages. The variety of linkages associated with
the top seven metropolitan areas is yet further evidence of their key role as central
switching centers for flows of information on the Internet.
The most striking finding of this analysis is that among the top seven metropoli-
tan areas, the majority ofbackbone capacity was used to link these regions to each
other or to provide service within the region, rather than to connect less important
cities and outlying areas. This aspect of the emerging backbone network merits
further research as it suggests that these metropolitan areas do not serve as con-
duits for a national system of data distribution, but instead have coalesced into a
separate, highly networked urban system containing both the major producers and
most consumers of Internet services. This configuration is very similar to observa-
tions about so-called 'global cities' like New York, London, and Tokyo, which
have become largely disconnected from their national economies while being in-
creasingly integrated with each other.
The backbone analysis was the most complete and accurate ofthese three analy-
ses because the data set involved was more limited and manageable with existing
analytic techniques. However, like the domain and host count measures, it should
serve only as a general indicator of relative differences in magnitude of Internet
accessibility and adoption among cities, regions, and metropolitan areas. Notwith-
standing, the stark difference in backbone capacity between the top seven metro-
politan areas and the rest of the nation was the clearest conclusion drawn from this
series of studies.
The Role ofthe Real City in Cyberspace 183
10.4 Conclusions
The research on the geography of the Internet summarized in this paper was con-
ducted over a two-year period from 1996 to 1998. Despite substantial difficulty in
obtaining accurate, timely, and comprehensive data sets, our results indicate sev-
eral important trends, as weIl as questions for future research. When compared,
the three sets of data indicate a consistent metropolitan dominance of the Internet
in the United States. This seetion summarizes our collective findings and observa-
tions.
Most significantly, the results of this research demonstrate conclusively that a
select group of metropolitan areas, and in particular their central cities, over-
whelmingly dominate the Internet in the United States. Above and beyond their
status as centers of population and employment, these regions consistently lead the
nation in the magnitude, density, and growth of Internet clusters. This fact calls
into question many assumptions regarding the spatial effects of information tech-
nology and telecommunications, as many influential authors have promulgated a
deterministic view that new technologies will lead to a radical decentralization of
population and economic activity (Gilder 1995, Negroponte 1995, Toftler 1980).
This view is not borne out by our research. The New York, San Francisco, Wash-
ington, and Los Angeles metropolitan areas that are the largest clusters of Internet
activity appear to be employing these technologies to reassert their economic im-
portance in the American urban system. Furthermore, the presence of a set of
smaller, Internet-savvy metropolitan areas, including Austin, Boston, Miami, and
Seattle, suggest that rather than causing decentralization, the Internet may permit
the development of a more complex, networked urban system.
These studies also indicate the lingering importance geographicallocation in de-
termining accessibility to advanced telecommunications services. Cities like At-
lanta, Chicago, and Dallas are extremely important hubs for national Internet
backbone networks out of proportion to their share of other measures of Internet
activity. We infer that their geographie centrality is an important factor in this al-
location of network capacity. As a result, these cities have access to more, faster,
and less expensive Internet infrastructure than cities of comparable size that are
less ideally located. However, the absence of comparable levels of Internet devel-
opment in other centrally located cities, such as Detroit and Philadelphia, lead us
to believe that a certain set of socioeconomic prerequisites influence the allocation
ofthese technologies.
In our effort to measure and describe the emerging telecommunications land-
scape of the United States, we anticipated being able to measure flows of informa-
tion across the Internet to gain a greater understanding of which regions were net
information producers, and which were net information consumers. Abler (1970)
and Mitchelson and Wheeler (1994) used this methodology in the past to deter-
mine the relative position of cities within an urban hierarchy. While we were un-
able to gather such data, future research may benefit from a proposal to integrate
geographie information in the Internet's domain name system. The Request For
184 M.L. Moss and A.M. Townsend
Comment 1976 proposal (RFC 1976), Putting Locations in the Domain Name Sys-
tem, would theoretically permit researchers to measure the flow of information
across monitoring stations along the Internet in real time and to more accurately
determine the location of host computers than is currently possible (Davis 1998).
Perhaps the most important conclusion we have made is that researchers study-
ing the Internet need to be very sensitive to the technical intricacies of the systems
they investigate. The relevance of the host count measurement was seriously un-
dermined by several factors discussed earlier, such as the use of firewalls to mask
corporate networks. Without a clear understanding of exactiy how the Internet is
constructed, we probably would have accepted those results beyond their actual
significance. Furthermore, because much of this infrastructure is actually an array
of intangible data and logical constructs (domains, virtual backbones) or easily
reconfigured electronic equipment (host computers and fiber optic networks),
means it can be reallocated almost instantly in response to market shifts, natural
disasters, etc. By definition, the Internet is highly volatile and in constant flux.
Research runs the risk of being an anachronism before it is ever published. For
cities, this means that unlike the televised urban riots that launched Lyndon John-
son's War on Poverty or the graphie images of Rust Belt cities in decay that
helped put Ronald Reagan in the White House, Internet communities can disap-
pear with the flip of a switch.
Finally, there is a strong need for systematic data gathering on the geography of
Internet development and information flows. The RFC 1976 proposal is a good
candidate, and we hope its key points will be incorporated in the new version of
the basic Internet Protocol, IPv6, now being designed.
One of the many interesting ideas to emerge from the NCGIA Project Varenius
specialist meeting, from which the chapters in this volume are drawn, was the
concept of a federally funded Internet Census. Historically, Congress has often
charged the Census Bureau to conduct exhaustive surveys on the emerging tele-
communications industries. A full twelve pages of the 1852 economic Census
were dedicated to the telegraph (Standage 1998), and two special studies of the
telephone industry in 1902 and 1907 each contained over 500 pages of statistics
about the structure and geography of the fledgling industry. These important
documents were very influential in highlighting the uneven geographie distribu-
tion of telecommunications capabilities and paved the way for the universal ser-
vice provisions that governed most national telecommunications policies for the
remainder ofthe 20 th century.
As this chapter has shown, the Internet in the United States is spreading in a
highly uneven fashion, strongly favoring profitable markets and punishing eco-
nomically distressed cities and remote areas. If the Internet will be the primary
vehicle for commerce and communications over the next 50 years, as many in the
Clinton-Gore Administration believe, accurate data about the location of Internet
infrastructure and users will be essential for making sound policy decisions. Public
officials have chosen to defer to the market in regulating the development of the
Internet, but this research shows that a laissez-faire approach has not led to an
equitable distribution of access and technology among and within cities and met-
The Role ofthe Real City in Cyberspace 185
References
Abler, R 1970. What makes cities important? Bell Telephone Magazine.
Castells, M. 1989. The Information City: Information Technology, Economic Restructuring,
and the Urban-Regional Process. Cambridge MA: Blackwell.
Davis, C. 1998 RFC 1976 Resources: Putting Locations in the Domain Name System. E1ec-
tronic document [http://www.kei.comlhomepages/ckd/dns-Ioc/]. Viewed 28 July 1998.
Dodge, M. and Shiode, N. 2000. Where on the Earth is the Internet?: An empirical investi-
gation ofthe spatial patterns ofInternet 'real estate' in realtion to geospace in the United
Kingdom. In Wheeler, lO., Aoyama, Y. and Warf, B. (eds.). Cities in the Communica-
tions Age: The Fracturing ofGeographies. New York: Routledge.
Garreau, J. 1991. Edge City: Life on the New Frontier. New York: Doubleday.
Gaspar, land G1aeser, E. 1996. Information technology and the future of cities. National
Bureau ofEconomic Reseach Working Paper, No. 5562. Cambridge MA.
Gilder, G. 1995. Forbes ASAP, 27 February, p.56.
Gottmann, l 1961 Megalopolis. Cambridge MA: MIT Press.
Greenstein, S., Lizardo, M. and Spiller, P. 1997.The evolution of the distribution of ad-
vanced large scale information infrastructure the United States. National Bureau of Eco-
nomic Research Working Paper, No. 5929. Cambridge MA.
Hall, P. 1997. Modeling the post-industrial city. Futures 29: No. 4/5
Kolko, J. 1998. Personal communication. l Kolko can be reached at email:
kolko@nber.org
Matrix Information and Directory Services, Inc. 1996. Matrix Maps Quarterly, No. 303
(January). Austin TX.
Mitchelson, RL., and Wheeler, lO. 1994. The flow of information in agiobai economy:
the role of the American urban system in 1990. Annals of the Association of American
Geographers, 84:87-107.
Moss, M.L. and Townsend, A.M. 1996. Leaders and losers on the Internet. Taub Urban
Research Center, New York University, September
[http://urban.nyu.edu/researchll-and-I/report.html]
Moss, M.L. and Townsend, A.M. 1997a. Manhattan leads the 'net nation. Taub Urban Re-
search Center, New York University, August.
[http://urban.nyu.edu/researchldomains/domains.html]
Moss, M.L. and Townsend, A.M. I 997b. Tracking the 'net: Vsing domain names to meas-
ure the growth ofthe Internet in U.S. cities. Journal ofUrban Technology. 4: No. 3.
Moss, M.L., and Townsend, A.M. 1998. Spatial analysis ofthe Internet in V.S. cities and
states. Paper presented at the Urban Futures - Technological Futures Conference, Dur-
harn UK, 23-25 April. [http://urban.nyu.edu/researchlnewcastle/newcastle.html]
Negroponte, N. 1995. Being Digital. New York: Knopf.
Nunn, S. and Warren, R. 1997. Metropolitan telematics infrastructure and capacity for eco-
nomic development in the information society. Center for Urban Policy and the Envi-
ronment, Indiana University.
Pred, A.R 1973. Urban Growth and the Circulation ofInformation. Cambridge MA: Har-
yard University Press.
186 M.L. Moss and A.M. Townsend
Rickard, J. 1997. The Internet - what is it? Boardwatch Magazine - Internet Service Pro-
viders Quarterly Directory. Littleton CO, Fall.
Standage, T. 1998. The Victorian Internet. New York: Walker and Co.
Sui, D.Z., and Wheeler, J. 1993. The location of office space in the metropolitan service
economy of the United States, 1985-1990, The Professional Geographer 45 :33-43.
Toffler, A. 1980. The Third Wave. New York: William Morrow & Co.
11 Accessibility to Information within the
Internet: How can it be Measured and
Mapped?
Martin Dodge
Centre for Advanced Spatial Analysis, University College London, 1-19 Torrington Place,
Gower Street, London, WClE 6BT, UK. Email: M.Dodge@ucl.ac.uk
11.1 Introduction
One definition of the Internet is ' ... a eolleetion of resourees that ean be reaehed
from those networks' (Krol and Hoffman 1993, 1). This definition provides the
starting point for my eoneeptualization of aeeessibility in the Information Age. I
will examine how one ean begin to measure and visualize the aspeets of aeeessi-
bility to information resourees within the Internet. My diseussion starts with the
assumption that a person has physical aeeess to the Internet, via a networked eom-
puter 1• Onee this 'physieal' eonneetivity has been overeome, how aeeessible are
the information resourees, people, and eleetronie plaees available online? What
are the future aeeessibility issues that need to be eonsidered to realize the full po-
tential of the Internet beyond basic eonneetivity? As the Microsoft mantra says,
where do you want to go today?, so what resourees are aeeessible within the Inter-
net and how do you reaeh them? This is very mueh a concern with individ-
ual accessibility and with developing a behavioral geography for Cyber-
space (Brunn 1998, Kwan 1998).
It is also important to eonsider how these dimensions of aeeessibility ean be
represented, particularly given the abstract nature of information spaces, which
often do not have natural spatial struetures. The online activities of searehing,
browsing and communicating in Cyberspace are ethereal and non-tangible to con-
ventional mapping techniques. Future representations need to be dynamic and
interactive, as weil as being readily available while navigating the Internet (De-
cember 1995, Dodge 1999b).
I will argue that the scope of geographical accessibility needs to be expanded to
encompass notions of information accessibility. The growing importance of the
Internet, and its layered services, for receiving and distributing all manner of in-
formation and for personal interaction, will require us to consider how concepts of
1 This is, of course, a big assumption and one that requires serious investigation; see, for
example, Wresch (1996) and Holdemess (1998).
188 M. Dodge
accessibility are played out within these electronic spaces. Gaining access in a
timely fashion to the right information resource, be it a Web page, an email, a
video clip, achat room, or a particular place in a virtual world, is problematic for a
number of human and technical reasons. The Web may weil facilitate easy access
to vast arrays of information from servers around the world, but this does not
mean one can find useful, current, reliable, and affordable information at the right
time. As Pirolli, Pitkow, and Rao (1996, I) comment, 'The apparent ease with
which users can click from page to page on the World-Wide Web belies the real
difficulty of understanding the what and where of available information'. The
Web is a vast array of information, but the ratio of noise to useful information can
be very high. The problems of information retrieval through searching and brows-
ing this massive space are becoming important for conceptualizing accessibility in
the Information Age. There is an increasing awareness of the problem of informa-
tion overload, with 'a tsunami of data crashing onto the beaches of the civilized
world' (Wurman 1997, 15). Accessibility to too much information is potentially as
significant an issue as accessibility to too little information. Excessive information
impedes its assimilation and therefore does little to improve knowledge and un-
derstanding (Shenk 1997). A great deal of effort is being directed by researchers in
a range of disciplines to cope with the problem of information retrieval and infor-
mation overload through filtering, structuring, analyzing and visualizing informa-
tion to aid the limited human capacity to search for, absorb, and comprehend in-
formation (Berghel 1997, ASIS 1998). Much of this research is relevant to
broadening the scope of geographie accessibility to encompass information spaces
of the global Internet.
In this chapter I discuss how we can develop the theme of informa-
tion accessibility, examining the following topics (1) nature of the dif-
ferent Internet information space, (2) issues of network performance
and tools to try and diagnose problems, (3) the importance of search
engines and the problem of searchability of resources, and (4) accessi-
bility problems caused by the design of information spaces. The issues
of measurement and representation of information accessibility are
highlighted in detail for two particular Internet information spaces; an
empirical investigation of accessibility between Web sites using the
structural information contained in hyperlinks and an examination of
accessibility in a 3d virtual world on the Internet.
It is important to be aware that the Internet provides users with a number of dis-
tinct services, which are often thought of as different spaces, with differing virtual
landscapes. In particular, these different spaces support different types of informa-
tion exchange, degrees of synchronicity, and levels of social interaction. There-
Accessibility to Information within the Internet 189
fore, they are likely to require different measures of accessibility and forms of
graphie representation to appropriately model their true nature. At a fundamental
level, the different information spaces are caused by the different network proto-
cols used by software applications to communicate over the Internet, which give
rise to the different forms and functions apparent to the end-user. Figure 11.1
shows a sketch map produced by John December, showing the principal informa-
tion spaces and some of the connections between them (December 1995).
------
CyberMap Landmarks 03 DK \994
. . Searcher or 'nd.~
_ orLlnk
Sub)ectTree
qjII Root LI.t
Figure 11.1. Information spaces of the Internet - circa 1994. Reproduced with permission
from John December (1995).
largely unseen from the outside and so are difficult to quantify and map. More
recent work on conceptualizing the form and structure of different information
spaces includes Michael Batty's (1997) examination of Virtual Geography, Paul
Adams's (1998) discussion on Network Topologies and Virtual Place, Manuel
Castells's (1996) Space 01 Flows, and Brian Gaines's research on the human-
factors oflnternet information spaces (Gaines, Chen, and Shaw 1997).
There are a number of important issues that need to be factored into future com-
prehensive models of accessibility to information in networked space. I will dis-
cuss the following four issues:
• network performance,
• size of the information spaces,
• information findability and persistence, and
• information structure, design and user behavior.
A common joke is that the WWW really stands for the World-Wide Wait be-
cause of the poor and unpredictable performance of the Internet, particularly per-
ceived by the Web user waiting for a page to download. It is often asserted that
'distance is dead' on the Internet because it does not matter where the information
is geographically located in the world that the user is trying to access - it is only a
mouse click away. Instead, what really matters is where the site is located in tem-
poral space. Perceived response time replaces geographic distance as the key vari-
able in access to interactive information spaces like the Web. A slow response
time from a Web site that is physically nearby means that it will be perceived as
being more remote and inaccessible than a fast site that is thousands of miles away
on another continent. Limited human patience with interactive computer interfaces
means that response time is critical and even delays of a few seconds can prove so
frustrating that people simply give up and try to locate an alternative source closer
in temporal space (Nielsen 2000). If network performance degrades below certain
thresholds, so me information spaces on the Internet become effectively infinitely
distant and inaccessible because people will not bother to wait. The delays caused
by network performance also have financial implications for those people paying
by the minute for their access to the Internet. The complexity and self-organizing
nature of the underlying Internet infrastructure makes it difficult to determine
where performance problems that effect information accessibility are located in
the network (Huberman and Lukose 1997). Problems can be at any point in the
chain from the user's computer through the network links and nodes, to the target
server. Examining Internet traffic-flows reveals the incredible complexity of
Accessibility to Infonnation within the Internet 191
routes that data travels through the network, often traversing fifteen or more nodes
and crossing physical infrastructure owned and operated by competing Internet
service providers and telecommunications companies. It is possible to explore
Internet trafiic routing using utilities called traceroutes (Rickard 1996) and it can
be surprising just how complicated things are 'under the hood', so to speak. It is
amazing that the Internet works as well as it does! To take an exarnple, my Web
site is located at a commercial hosting service that is geographically about 1.5
kilometers from my office in University College London (UCL) and yet a trac-
eroute reveals that traffic takes twelve hops (different network nodes) to travel this
distance. Whereas traffic to the mirror site that is physically located in Washing-
ton DC, around 3,700 km from UCL only takes slightly longer, in network terms,
at fourteen hops. For measuring accessibility through the routing topology of the
Internet, these two Web site locations are pretty much equally distant from UCL.
However, there are some interesting variations in response times from these two
sites. The site in London is gene rally equally responsive throughout the day, how-
ever the accessibility response time to the mirror site in Washington depends on
the time of day. The response degrades noticeably in the afternoon when the great
mass of North American Internet users wake up and log on. It is well known that
for European Web surfers, the performance of the Web degrades significantly in
the afternoon when America wakes up. Clearly, measuring accessibility to global
information resources like the Internet and the Web will require the time
dimension to be fully integrated (Harvey and Macnab explore this theme in Chap-
ter 9).
At the network performance level, information accessibility would be measured
by three key parameters, (1) delays, known as network latency, (2) deliverability,
the problem of data being lost in transit and having to be resent, and (3) availabil-
ity of the network and servers, assessed by the amount of down-time. Much of the
research into the performance of Internet infrastructure is highly technical (e.g.,
Paxson 1997), however there is some research particularly relevant to the issue of
geographical accessibility. This inc1udes the work of John Quarterman who under-
takes real-time monitoring of Internet performance by ca\culating latencies to sev-
eral thousand sampie nodes in the global Internet from his headquarters in Texas
(Quarterman, Smoot, and Gretchen 1994, Quarterman 1997). The results are pre-
sented as animated maps that he terms Internet Weather Reports (lWR). Figure
11.2 shows one ofthe animation frarnes for California (Quarterman 1997; see also
http://www.mids.org/weather/). Quarterman claims that his analysis shows that
there has been a thirty- percent improvement in mean Internet latencies since he
started monitoring in 1994. There is also a study by Keynote Systems / Board-
watch into Internet backbone performance in the United States. This involves
large-scale testing of the performance of ten different networks from twenty-seven
different sampie points in different cities (Rickard 1997). Shane Murnion is also
undertaking research into the geography of Internet latencies, examining informa-
tion flows from UK academic Web servers to sixty-six countries (Murnion and
Healey 1998; see also Chapter 12). His findings show the existence of a distance-
decay effect in Web server audiences with increasing latency.
192 M. Dodge
5,000 ).< 64
3,000 ,.. 32
2,000 ).< 16
1,000 02:00AM ).< 8
500 PWT 4
300
200
100 8~
llllCnc:id doma.i.o~
Q:,p,7i,p:.(c) 1994
+10612..4el1 ·7eC2 t31 +1-512-4/52-01 Z7
".1)$
A....:ia, T,,-.. USA
--
hl1pi l'Mm'.mIdli.OIg mtls:jJ)mkl:li.olg
·010 1:5,000,000 A bel15Cau; EqLS1-ArOlal plOjIIc!Ion 98.0828_04:47 (0 2IJ 0)
Figure 11.2. Internet Weather Report (IWR) for California. Reproduced with permission
from John S. Quartermann, Matrix Information and Directory Services.
http://www.mids.orgl.
Accessibility to Infonnation within the Internet 193
Another key issue for reconfiguring accessibility measures for the information
age is to develop scaleable models to cope with the size, diversity, and dynamic
nature of Internet information spaces. Pinning down definitive figures on the size
of Internet information spaces from the cyber-hype and Net boosterism can be
difficult, but it is big and, more significantly, it is the fastest growing medium of
information and communication in history (ITU 1997, Clemente 1998)2. Recent
statistics show there were an estimated 36.7 million Internet hosts (Network Wiz-
ards 1998) and approximately 153 million Internet users worldwide (NUA 1999)
in January 1999. For the information resources potentially accessibility on the
Web, estimates from 1998 were that 2.8 million sites contain around 300 million
pages (Bharat and Broder 1998, Lawrence and Giles 1998). More importantly,
these studies reveal how incomplete the databases of even the largest search en-
gines are, only indexing at best a third of the Web. The consequence is that the
Web pages people are seeking may exist, but if they are not in the search engine
indexes they are invisible and, therefore, inaccessible because people will never
find them. Even if they are in the index, they may come so far down the list of
search results returned that most users will never see them, making them effec-
tively inaccessible. This has been termed the searchability factor by Esther Dyson
and is crucial in realistically modeling information accessibility (Dyson 1999).
Search engines have become the key access points - known as portals in the
current terminology - to information spaces on the Internet. They are rich indexes
of socially constructed information and, as such, offer a potentially useful data-
base for geographical research. The potential is beginning to be realized, for ex-
ample in research into the reproduction of the concepts of place onIine (Alderman
and Good 1997, J ackson and Purcell 1997, Alderman 1998, Henkel 1998, N orris
1998).
Another serious issue with modeling information accessibility is the problem of
persistence, or rather lack of it, on the Internet. The Internet and the Web are
changing every day, with information resources, sites, and virtual places appearing
and disappearing. A flick of a switch or a press of a delete key can cause whole
parts of an information space to simply disappear without a trace. Information
structures in Cyberspace are much less permanent than those of the real world.
Can accessibility measures cope with this?
The actual design of the information space can have an impact on its accessibil-
ity (Nielsen 2000). Like real-world cities and buildings, poorly designed Web sites
and pages are rendered inaccessible to certain groups of users. The virtual world
of information certainly suffers from the same degree of bad architecture, as does
the material world (Wurman 1997). We have all seen Web sites with poor choice
of fonts, colors, or frames, for example, which make them practically unusable.
The need for accessible design is especially important for visually impaired people
who surf with text-only software. There was arecent high profile case reported in
2 Although the Internet is a large infonnation space, it is still small in absolute volume
tenns compared to other infonnation domains, particularly broadcast television. See the
paper by Lesk (1998) for a fascinating examination ofthe sizes of different infonnation
domains.
194 M. Dodge
the United Kingdom where the newly re-vamped Web site for Number 10 Down-
ing Street was so badly designed that it was said to be 'staggeringly inaccessible'
(Jellinek 1998). Work is ongoing to improve the physical design and accessibility
of Web sites, coordinated through the Web Accessibility Initiative
(http://www.w3.org/WAII) of the World Wide Web Consortium. However, the
problem is difficult to resolve, as many Web-site designers are keen to use the
very latest technologies, which can easily make their sites inaccessible to many
average users. A survey of UK Web sites in 1997 found that only 30 percent of
pages were completely accessible to all users (Beckett 1997).
To develop new models of infonnation accessibility we need to know about the
content and structure of infonnation spaces and how users behave in them. There
is some useful work trying to answer these questions, but our quantitative knowl-
edge of Cyberspace is far from complete, particularly compared to our knowledge
of real-world spaces (Fagrell and S0rensen 1997). There are several reasons for
this, including the sheer newness of some of the infonnation spaces and their in-
visibility to the conventional monitoring and census-taking methods developed for
the material world (Batty 1990). Governments have not, until very recently any-
way, realized the significance of the infonnation spaces and so have made no at-
tempts to gather statistics on them. However, there are a number of interesting
academic studies that have begun to fill in the blanks. For contents of the Web,
there is the work ofBray (1996), Woodruff, et al. (1996) and Fagrell and S0rensen
(1997). The structure of the infonnation spaces, in particular the Web, has been
examined; see the exemplary work of James E. Pitkow (Pirolli, Pitkow and Rao
1996, Pitkow 1998) and Bray (1996). Researchers are beginning to analyze and
model how users behave in Cyberspace, as seen in the work by Hubennan, et al.
(1998) who have devised a law ofsurfing to describe user behavior.
My colleague Naru Shiode and I are researching the spatial structure ofthe Web
by analyzing the hyperlinks between sites (Dodge 1998). The data on how Web
sites are linked together can be used to model accessibility for this particular in-
fonnation space. In a preliminary investigation, we analyzed a smalI, manageable
sub set of the Web, the site of the major universities and colleges in the United
Kingdom, some 122 nodes. We used the AltaVista search engine
(http://www.altavista.com/) to gather statistics on the size of each Web site (de-
fined by the number of pages) and the number of hyper links between them, which
took 14,884 separate queries to AltaVista. The results showed that the 122 sites
contained more than one million Web pages and over 450,000 hyperlinks (al-
though the vast majority were intemallinks within individual sites).
Accessibility to Infonnation within the Internet 195
20 Links
The data on the hyperlink connectivity between sites were analyzed to deter-
mine the most accessible Web site. To do this, the distance from each university
to every other one was calculated. In the Web, virtual distance was calculated as
inversely proportional to the number of hyperlink connections between two points.
In the example shown in Figure 11.3, site B is much closer in virtual distance to
site A than is C. The Web site that had the lowest average distance, i.e., was clos-
est to all the others, was designated the central, most accessible one.
Figure 11.4. Web Scan showing the most accessible Web sites of UK universities
(Source: Dodge 1998)
196 M. Dodge
The virtual distances between all 122 sites were measured and stored as a large
graph where the edge lengths were assigned the distance value. The graph was
analyzed to find the shortest-path distance from each university to every other one
using the Dijkstra algorithm. The mean of these shortest-path distances was calcu-
lated for each Web site and then normalized by dividing by the number of Web
pages to take account of the influence of variations in Web-site size. The resulting
ranking showed that the University of Oxford's Web site (http://www.ox.ac.uk)
had the smallest, normalized, mean shortest path distance to all other site; hence, it
was declared as the most accessible Web site in our experiment. The graph ofvir-
tual distance was then used to measure the shortest-path distance from Oxford to
each university and this value was used as a metric of accessibility in Web space
and was called the WebX distance.
To begin to understand the structure and differential accessibility of academic
Web sites, as measured by the WebX distance, it was necessary to visualize the
position of sites in relation to Oxford. To achieve this we used a radar-type map
called a Web Scan. In the Web Scan, Oxford becomes the central point of gravity,
around which planetary Web sites rotate, their orbital distances being equal to
their WebX distance. Figure 11.4 shows a Web Scan for the most accessible Web
sites, those closest to Oxford. What is immediately striking are the two giant sites
very close to the Oxford center point. These are the University of Cambridge
(http://www.cam.ac.uk/) and University of Edinburgh (http://www.ed.ac.uk/).
which have large Web sites and are very closely interconnected. Cambridge has a
WebX distance of 10 and Edinburgh is only slightly further out at 13. Imperial
College (http://www.ic.ac.uk/) comes next with a WebX distance of 26, double
that of the second-place site. Further out from the top three, there is a cluster of
sites around the 40-WebX mark. These are Heriot-Watt University, the universi-
ties of Leeds, Glasgow, Southampton, Queen Mary and Westfield College, and
my own institution - University College London (http://www.ucl.ac.uk/). All
these universities are well connected and accessible in the academic Web. UCL
has a WebX score of 42, placing it in seventh place away from Oxford, a respect-
able place given its historic place in the development of Cyberspace in the United
Kingdom, as it was the first organization in Britain connected to ARPANET, the
Intemet's forerunner, back in 1973 (Salus 1995). There is, then, a slight gap until
the next Web sites are encountered, including large metropolitan universities such
as Birmingham, Liverpool, Bristol, Sheffield, and Manchester, and, as well,
smaller provincial institutions like Y ork and Stirling. Finally, right on the edge of
this scan is the University of Durham (http://www.dur.ac.uk/) with a WebX score
of exactly one hundred. Further results of our preliminary work are presented in
Dodge (1998). We are extending this type ofvirtual-network accessibility analysis
using a larger, more realistic sub set of the World-Wide Web. This type of analysis
of the structure of the Web, I believe, provides a potentially useful avenue for de-
veloping accessibility measures suitable for the Information Age.
Aeeessibility to Information within the Internet 197
Virtual worlds are a form of information space on the Internet that provides a
simulated environment in which multiple users can interact with each other in
real-time. Crucially, the users have a bodily representation in the space as an ava-
tar, and the simulated environment is graphically rendered in 2d, 2.5d or full 3d. A
number of virtual worlds from competing companies have emerged on the Internet
since mid-1995 (Rossney 1996, Damer 1998). They are used by many thousands
of people who have constructed new forms of social interaction and a distinct
sense of community and cultural identity to suit the unique characteristics of these
spaces (Rossney 1996, Donath 1997, Schroeder 1997, Damer 1998, Rafae!i, Sud-
weeks and McLaughlin 1998). In many respects the users can be said to inhabit
these worlds and have developed a real sense of place in Cyberspace. Virtual
worlds have pre-defined and programmed geographical dimensions, architectural
structures, and rules of avatar movement, they are truly information spaces (An-
ders 1998). They provide a fascinating new re alm, arguably at the cutting-edge of
the Information Age, in which to explore the meaning of geographical accessibil-
ity. The spatial nature of virtual rea!ity and Internet virtual worlds has, so far, re-
ceived !ittle attention from academic geographers, notable exceptions being Hillis
(1996) and Taylor (1997).
along the principal compass axes apparent in Figure 11.6. The spokes are clearly
evident in the December 1996 map, although in the second map, taken just over a
year later, they have become less pronounced as fill-in development has taken
place
The ability to teleport is a powerful feature, but interestingly it was not made
available to users when AlphaW orld was first launched in the summer of 1995. It
has only been progressively introduced for fear of its affects on the world. As the
AlphaWorld newspaper, the New Warld Times, reported in November 1995:
Teleportation! Yes Teleportation! The one most common request of AlphaWorld citi-
zens has been teleportation ... With teleportation more of AlphaWorid will become
readily accessible .... There is still some concern that teleportation will ruin the simu-
lation ofreality in AlphaWorld. In order to keep this simulation within bounds, tele-
portation will be implemented in a somewhat limited fashion. A 'Grand Central Teles-
tation' located at or near Ground Zero will enable citizens to teleport to key locations,
from which they can travel more easily to their destinations of choice. (New Warld
Times, #4, p. 2, http://vrnews.synergycorp.com/nwtJ).
AlphaW orld also warps conventional spatial norms and mIes of physical
movement at a local scale that have an impact on geographical accessibility. Y ou
are able to fly, unaided, above and even below the ground. To achieve this one
simply presses the + and - keys to effortless float up and down. It is also possible
to walk through any walls and structures by holding down the shift key, which has
the effect of making all objects immaterial to your avatar. These two god-like
pawers have had a significant impact on the architectural design of buildings in
AlphaWorld (Damer 1997). Visibility in AlphaWorld is also artificially con-
strained because it is only possible to see a maximum of 120 meters in any direc-
tion. This is due to the limits of graphics hardware and software on pes to render
a larger 3d landscape in real-time. However, the effect on the space is really quite
unnerving, like walking around in an opaque bubble 120 meters across, where
streets and buildings appear to end, with a sharp, artificial looking cut-off line.
This impacts on local accessibility because it is hard to orientate and navigate with
no fixed landmarks and distant vistas.
200 M. Dodge
Figure 11.6. Maps ofthe city at the center of AlphaWorid in December 1996 (top) and
February 1998 (bottom). Reproduced with permission from Vilett (1998).
Accessibility to Information within the Internet 20 I
11.6 Conclusions
In this paper I have been concerned with the idea of information accessibility
within the virtual spaces of the global Internet. Two particular Internet-based in-
formation spaces, the Web and Virtual Worlds, have been examined as exemplar.
One approach to measuring and mapping the relative accessibility of Web sites is
to use the structure of hyperlinks between them to calculate measures of virtual
distance. In contrast, virtual worlds provide a fascinating challenge to the conven-
tions of geographic accessibility. Although the worlds have a tangible, simulated
geographic environment that has many of the spatial characteristics of the real
world, they can also warp the physical conventions of distance and travel. If, as
some predict, these kinds of shared, 3d environments became prevalent as the next
generation of information interface, then it will be vital to understand how acces-
sibility effects the way people navigate and find places in the virtual worlds (An-
ders 1998).
In the Information Age the importance of access to information spaces, such as
the Web or places in a virtual world, will increasingly take precedence over access
to physical facilities in the real-world for certain important human activities. Be-
ing able to quantity and visualize accessibility to these virtual information spaces
will be an important challenge in extending the notions of geographic accessibility
to encompass Cyberspace.
References
Adams, P. 1998. Network topologies and virtual place. Annals ofthe Association ofAmeri-
can Geographers 88:88-106.
Alderman, D.H. and Good, D.B. 1997. Exploring the virtual South: The idea ofa distinctive
region on 'The Web'. Southeastern Geographer 37:20-45.
Alderman, D.H. 1998. Finding the heart ofDixie in cyberspace: The internet as a new labo-
ratory for cultural geography, paper presented at The Association of American Geogra-
phers 94th Annual Meeting, March, Boston MA.
Anders, P. 1998. Envisioning Cyberspace: Designing 3D Electronic Space. New
York:.McGraw HilI.
ASIS 1998. American Society for Information Science (ASIS) '98 - Information Access in
the Global Information Economy, 24-29 October, Pittsburgh PA.
Batty, M. 1990. Invisible cities, Environment and Planning B: Planning and Design
17: 127-30.
Batty, M. 1997. Virtual geography, Futures 29: 337-52.
Beckett D. 1997. 30% accessible <:> A survey of the UK Wide Web, Proceedings of the
Sixth International WWW Conference, April, Stanford CA.
Berghel H. 1997. Cyberspace 2000: Dealing with information overload, Communications of
the ACM, February 1997 402: 19-24.
Bharat, K. and Broder, A. 1998. A Technique for measuring the relative size and overlap of
Web search engines, Proceedings ofthe Seventh International WWWConference, April,
Brisbane Australia, 379-88.
202 M.Oodge
Bray, T. 1996. Measuring the Web, Computer Networks and ISDN Systems 2:8 7-II, 993-
1005.
Brunn, S.O. 1998. The Internet as 'the new world' of and for geography: speed, structures,
volumes, humility and civility, GeoJournal45: 1-2,5-15.
Cairncross, F. 1997. The Death of Distance: How the Communications Revolution Will
Change Dur Lives. Cambridge MA: Harvard Business School Press.
Castells, M. 1996. The Rise ofthe Network Society, Vol. I of The Information Age: Econ-
omy, Society and Culture. Oxford: Blackwell Publishers.
Clemente, P. 1998. The State ofthe Net: The New Frontier. New York: McGraw-Hill.
Couclelis, H. 1996. The death of distance, Environment and Planning B: Planning and
Design 23: 387-89.
Oamer, B. 1998. Avatar!: Exploring and Building Virtual Worlds on the Internet. Berkeley
CA: Peachpit Press.
Oecember, 1. 1995. A cybermap gazetteer. In Staple, G.C. (ed.) TeleGeography 1995:
Global Communications Traffic Statistics & Commentary. Washington OC: TeleGeogra-
phy Inc., 74-81.
Oodge, M. 1998. Journey to the centre of the Web. In Staple, G.C. (ed.) TeleGeography
1998/99: Global Communications Traffic Statistics & Commentary. Washington OC:
TeleGeography, Inc., 150-54.
Oodge, M. 1999a. Explorations in AlphaWorld: The geography of3d virtual worlds on the
Internet, paper presented at RGS-IBG Annual Conference - Geographies of the Future,
Leicester UK, 4-7 January.
Oodge, M. 1999b. Atlas of Cyberspaces. London: Centre for Advanced Spatial Analysis,
University College London (http://www.cybergeography.org/atlas/).
Oonath, J.S. 1997. Inhabiting the virtual city: The Design ofSocial Environmentsfor Elec-
tronic Communities. PhO thesis, MIT (http://judith.www.media.edu/Thesis/).
Oyson, E. 1999. Search and searchability, The Guardian, Online section, 14 January, 11.
Fagrell, H. and Sorensen, C. 1997. !t's life Jim, but not as we know it!, Proceedings ofthe
WebNet '97 Conference, October, Toronto. (http://ifi.uio.no/iris20/proceedingsI35.htm).
Gaines, B.R., Chen, L.L. and Shaw, L. 1997. Modeling the human factors of scholarly
communities support through the Internet and World Wide Web, Journal ofthe American
Society for Information Science 48:987-1003.
Henkel, T. 1998. An Internet artifact: place-concepts revealed in the virtual world. Paper
presented at the Telecommunications and the City Conference, Athens GA, University of
Georgia, March.
Hillis, K. 1996. A geography of the eye: the technologies of virtual reality. In Shield, R.
(ed.) Cultures of the Internet: Virtual Spaces, Real Histories, Living Bodies London:
Sage Publications, 70-98.
Holderness, M. 1998. Who are the world's information-poor? In Loader, B. (ed.) Cyber-
space Divide: Equality, Agency and Policy in the Information Society. London: Rout-
ledge, 35-56.
Huberman, B.A., Pirolli, P.L.T., Pitkow, 1.E. and Lukose, RJ. 1998. Strong regularities in
World Wide Web surfing. Science 280(3 April):95-7.
Huberman, B.A. and Lukose, R.M. 1997. Social dilemmas and Internet congestion. Science
277 (25 July):535-7.
ITU 1997. Challenges to the Network: Telecommunications and the Internet. Geneva: In-
ternational Telecommunication Union, September.
Accessibility to Infonnation within the Internet 203
Jackson, M.H. and Purcell, D. 1997. Politics and media richness in the World Wide Web
representation ofthe fonner Yugoslavia. The Geographical Review 87:219-39.
Jellinek, D. 1998. Door policy at Number 10, The Guardian Online section 7 May, 7.
Krol, E. and Hoffman, E. 1993. FYI on 'what is the Internet?' Requestfor Comments: 1462
(May 1993), (ftp://ftp.ripe.netlrfc/rfcI462.txt).
Lawrence, S. and Giles, C.L. 1998. Searching the World Wide Web, Science 280(3
April):98-100.
Lesk, M. 1998. How much infonnation is there in the world? Paper presented at the Time &
Bits: Managing Digital Continuity conference, Getty Infonnation Institute, February, Los
Angeles (http://www.gii.getty.edu/timeandbits/ksg.html).
Murnion, S. and Healey, R.G. 1998. Modelling distance decay etfects in Web server infor-
mation flows, Geographical Analysis 30:285-303.
Network Wizards 1998. Internet Domain Survey
(http://www.nw.comlzone/WWW/top.html).
Nielsen, J. 2000. Designing Web Usability: The Practice of Simplicity. Indianapolis: New
Riders Publishing.
Norris, D.A. 1998. The developing world and the World-Wide Web: Search engine topical
coverage, 1997. Paper presented at The Association of American Geographers 94th An-
nu al Meeting, March, Boston MA.
NUA 1999. How many online? (http://www.nua.ie/surveys/how_manLonline/index.html).
Paxson, V., End-to-end routing behavior in the Internet, IEEE/ACM Transactions on Net-
working, October 19975(5):601-15.
Pirolli, P., Pitkow, J. and Rao, R. 1996, Silk from a sow's ear: extracting usable structures
from the Web, Computer-Human Interaction (CH/'96) Conference, April, Vancouver,
Canada.
Pitkow, J.E. 1998. Summary of WWW characterizations. Paper presented at The Seventh
International World-Wide Web Conference, April, Brisbane, Australia.
Quartennan, J.S., Smoot, C.M. and Gretchen, P. 1994. Internet interaction pinged and
mapped, Proceedings ofthe INET'94 Conference, Internet Society, July, Prague, 522:1-4.
Quartennan, J.S. 1997. The Internet weather report. In Staple, G.C. (ed.) TeleGeography
1997/98: Global Communications Traffic Statistics & Commentary. Washington De:
TeleGeography Inc., 69-72.
Rafaeli, S., Sudweeks, F. and McLaughlin, M. 1998. Network & Netplay: Virtual Groups
on the Internet. Cambridge MA: MIT Press.
Rickard,1. 1996. Mapping the Internet with Traceroute. Boardwatch Magazine 10 (Decem-
ber). (http://boardwatch.internet.comlmag/96/declbwm38.html).
Rickard, J. 1997. Backbone perfonnance measurement, Boardwatch Magazine Directory of
Internet Service Providers 2(Fall):22-31.
Rossney, R. 1996. Metaworlds. Wired 4.06: 140ff.
Salus, P.H. 1995. Casting the Net : From ARPANET to Internet and Beyond ... Addison-
Wesley Publishing Company.
Schroeder, R. 1997. Networked worlds: social aspects of multi-user virtual reality, Socio-
logical Research Online 2:4. (http://www.socresonline.org.uklsocresonline/2/4/5.html).
Shenk, 0.1997. Data Smog: Surviving the Information Glut. San Francisco: Harper.
Taylor, 1. 1997. The emerging geographies of virtual worlds, The Geographical Review
87(2): 172-92.
Vevo, 1999. AlphaWorld Mapper (http://awmap.vevo.coml).
204 M. Dodge
Shane Murnion
connection between those nations and the United Kingdom. The strength of the
relationship is both c1ear and striking.
80
I::
•
•
0
70
.. •
~
:,.
'5
c. 60
0
c. •
-
= ,.~ ~
50
....
=
40
,.
VI
CI.)
I::
0
.c 30
•• • •
•• • •
c.
.!!
20
••
.. •"•••• •••
CI.)
I-
•• •
..
ci
z 10
•
0 •
0 1000 2000 3000 4000 5000
Average National Latency from UK
Figure 12.1. Telephone availability for various countries vs. the speed ofthe Internet con-
nection (measured using latency) between those countries and the United Kingdom.
40000.-----------------------------------------,
•
-
30000
• •
..
l'IiI
·l'IiIa ••
•• • •
...
:'
u
CI)
D-
D-
a
20000
.-
CI
••••
#
10000
• • •
• • • •
o~--·-'--·-~r~~4f·~·-·-·~~~~·~·-·--~·~~..~~------~
o 1000 2000 3000 4000 5000
Average National Latencyfrom UK
Figure 12.2. Gross Domestic Product (GDP) for various countries versus speed of the
Internet connections (measured using latency) between those countries and the United
Kingdom.
Spatial Interaction Models ofInformation Flows 207
Thus there are two major areas where the growth of e-commerce will have a
significant impact - on the developed economies that rely on the quaternary busi-
ness sector and on developing nations. The effects on the first area are difficult to
predict, however some insights into what may happen to developing nations may
be drawn from previous studies in communication geography. Certainly the ef-
fects of the Internet on developing economies is starting to attract the attention of
development consultants (Daly 1999). One study of telecommunication network
growth from urban to rural regions in the United States and elsewhere showed that
businesses based in the developed urban centers tend to dominate over rural-based
business (Abler 1991). This pattern is likely to repeat itself in the relationship be-
tween businesses in the developed and developing regions. A further area of con-
cern for developing regions is the asymmetry of latency space. Previous studies
show that there is a connection between the length of time required to access
Internet information and the amount of information accessed (Murnion and Healey
1998). Consider the hypothetical situation of a region A that has a poor Internet-
connected network exhibiting high latency values. For customers within the region
all Internet services delivered from sites on the local network will appear far away
in latency space. If the network borders a highly developed network from a devel-
oped region B with very low latency values then services based in B will appear to
be almost as close as their own internat services. Furthermore with the enhanced
levels of Internet expertise and business experience available within developed
regions, B's services may weil be superior in quality to A's. As a result customers
will tend to import services from Brather than use their own region's services.
However, for customers within the B region any service originating from A will
seem much farther away in latency space and of poorer quality than their own in-
ternal services and, as such, they are unlikely to use them. Thus it seems that the
inequalities between developed and developing regions will be exacerbated by
Internet commerce. To determine the effects of Internet services and the links be-
tween demand for these services and network quality as it varies regionally, we
require so me method of determining the quality of the network between two re-
mote points. In this chapter we discuss previous attempts at determining the link
between network quality and demand and examine how the methodologies used
may be extended to facilitate the type of analysis required.
One method that obviously applies in the analysis of Internet information flows
and their effects is that of spatial interaction modelling and indeed this method has
been used in a closely related study of telecom information flows in Europe
(Fischer and GopaI1998). However spatial interaction modelling can only be ap-
plied if space has some discernible effect on the flows studied. From an examina-
ti on of possible metrics and measuring tools, Murnion (2000) suggested that la-
tency, as measured using a utility called ping, might prove suitable in this type of
208 S. Mumion
analysis. Using this method Murnion and Healey (1998) undertook an analysis of
information flowing from UK academic Web servers to the rest ofthe world. The
aim of the study was to use a simple gravity model to determine whether or not
information flows on the Internet were affected by space. The study showed that
audiences accessing information from WWW servers are mostly located near to
those Web servers in latency space. The main result from the analysis was that the
catchment areas for WWW servers are regional in nature rather than global. Fig-
ure 12.3 illustrates the decay curve extracted from the analysis. The result is intui-
tive in that one might expect users would tend to choose a WWW server that can
provide information rapidly over one that provides the same information more
slowly. Although useful, the scope ofthe analysis was limited in some respects. In
this work an attempt is made to expand upon this initial work.
- '~
'Cij
5
Figure 12,3. Latency distance decay curve derived in the Mumion and Healey (1998)
study.
The main limitation of the previous study was its scope. Only information flows
from one sector, that of university academic information servers were examined.
This type of service exists on a high-performance homogeneous network (JANET)
and the types of services supplied were also highly homogeneous. One might ex-
pect that the wide variety of commercial services available might exhibit different
modes of behavior. Furthermore the study only covered the reasonably simple
Spatial Interaction Models ofInformation Flows 209
1.2
•
c
0.8
• ••
••
0
1a
e.. 0.6
0
(.) •
04
•
0.2
•
0 •
0 200 400 600 800 1000 1200
avg latency
Figure 12.5. Correlation on ten routes between latency in each direction plotted against the
average latency in both directions for that route.
1------- --~~I
I 1200
I~ 1000
I
eI:~ 5
I
800 •
.~~ 600 I
I: ~
S:C
~
400
200 • I
o I
To attempt the latency triangulation, six computers were used, located in Santa
Barbara, Portsmouth, Cape Town, Perth, Helsinki, and Tokyo respectively. The
locations are shown in Figure 12.7.
Figure 12.7. The locations ofthe computers used in the Internet triangulation exercise
Figure 12.8. Simultaneous latency measurements taken for the attempted triangulation.
Spatial Interaction Models of Infonnation Flows 213
The computers at Santa Barbara, Portsmouth, Tokyo, and Perth were used as
latency range-finding stations. Simultaneous latency measurements were taken by
each of these stations to the computers in Cape Town and Helsinki as shown in
Figure 12.8.
As each of these latency measurements was gathered, the true latency between
the computer in Cape Town and the computer at Helsinki was also measured. The
measurements were taken every 15 minutes over aperiod of three days. The com-
plete data set was split into two parts. One third of the cases, randomly chosen,
were used to train a neural network, which attempted to predict the latency be-
tween Cape Town and Helsinki from the measurements taken by the triangulation
stations. The actual latency measurements for the training cases were used to de-
termine the error in the neural network training. The details of the creation of the
neural network are not included since the choice of modeling method is unlikely
to critically affect the result. The trained neural network was then used in an at-
tempt to predict the latency between Cape Town and Helsinki for the remaining,
as yet unseen, triangulation measurements. The results of the exercise are given in
the next section.
Figure 12.9 shows the measured latency against predicted latency for the unseen
measurements.
1000
900
~
I:
S 800
j
al 700
Ü
~
0.
600
500
400
400 500 600 700 800 900 1000
Actual latency
Figure 12.9. Latency predicted by the neural network against the actual measured
latency. The solid line represents a perfect result.
214 S. Mumion
Overall, the results are poor though statistically significant, giving an overall R2
of 0.27 on 600 measurements. However it is noticeable that there is one large clus-
ter of data points between 580ms and 610ms that contributes the majority of the
error observed. If the data points between 580ms and 610ms are filtered out then
the R2 value rises to 0.56. Figure 12.10 shows the actuallatency between South
Africa and Finland measured over time.
2500 r---------------------------------------,
2000
~ 1500
c:
GI
j 1000
500
o
Time
Figure 12.10. Latency between Helsinki and Cape Town (The time distance between each
ofthe large peaks is one day).
Examination of Figure 12.10 shows the curious result that the latency values
poody predicted by the neural network represent approximately the baseline
minimum latency values recorded. It may be that the triangulation method is better
at detecting changes in network traffic then in absolute values.
12.7 Discussion
assumptions as the commutation results show. The results of this study show that
the triangulation method holds some promise as a potential method for measuring
remote latencies. However the methodology needs further improvement to reduce
the errors in the technique such that it could be used as a practical monitoring tool.
The importance of Internet triangulation goes beyond simply providing a
method of measuring latency. For the method to work, it requires the location in
cyberspace of each of the objects measured. If it is possible to obtain a dataset that
contains the cyberspace location of a large number of objects for which geo-
graphical locations are known, then it may be possible to build a model that can
map between latency space and geographical space. Using such a model, it would
be possible to pinpoint in geographical space the location of a computer connected
to the Internet, simply by triangulating on its IP or Internet address. If this is pos-
sible then it should be feasible to build dynamic maps of Internet usage and activ-
ity, locating users and domains in geographical space. The possibility exists then
of building an Internet census that accurately reflects information use on agiobai
scale.
References
Abler, R.F. 1991. Hardware, software and brainware: mapping and understanding commu-
nication technologies. In Brunn, S.D. and Leinbach, T.R. (eds.) Collapsing Space and
Time: Geographical Aspects ofCommunications and Information London: HarperCollins
Academic.
Daly, lA. 1999. Measuring impacts ofthe Internet in the deve1oping. World. IMP Maga-
zine. http://www.cisp.org/imp/may99/daly/0599daly.htm..
Fischer, M., and Gopal, S. 1998. Artificial neural networks: A new approach to modelling
interregional telecommunication flows. In Haynes, K.E. et al. (eds) Regional Dynamies.
Cheltenham: Edward Elgar, 503-27.
JISC. 1998. Usage-related charges for the JANET network JISC Circular 3/98.
http://wwwjisc.ac.uklpub98/c398.htm.
MIDS. 1998. The Internet Weather Report. http://www.mids.org/weatherl
Murnion, S. 2000.Cyber-spatial analysis: appropriate methods and metrics for a new geog-
raphy. In Openshaw, S. and Abrahart R. (eds.) GeoComputation. The Netherlands:
Balkema Publishers.
Murnion, S., and Healey, R.G. 1998. Modelling distance decay effects in Web server in-
formation flows. Geographical Analysis 30(4):285-303.
University of Oregon. 1998. Internet Weather for News Hosts, http://twin.uoregon.eduliwr
13 Application of a CAD-based Accessibility
Model
Paul C. Adams
13.1 Introduction
'given' scale attached to the technology is derived from the rather obvious fact
that the information technologies are capable of supporting international commu-
nications. This scale, once accepted as real, seems all the more ominous since the
condition of glabality is not uniformly accessible. Those most able to use and di-
reet the system are the ones with the most money, education, and political power.
Thus, we see a global system dominated by powerful groups, accessible to power-
ful groups, and potentially destructive of individual agency and freedom (e.g.,
Castells 1996; 1997; 1998; Graham and Marvin 1996; Curry 1995). Without fully
analyzing our perspective, we have automatically applied a macro-scale lens to the
phenomenon. Its implications, political and otherwise, are in fact scale dependent
and look different at other scales.
To understand IT more cIearly, we must incIude individuallives in our analysis,
and particularly the routines people pursue in and through IT. Rather than simply
asking what people can do, we must ask what they da do. A portrait of extensible
individuals, though not self-sufficient, reveals meanings of IT that are missed in
aggregate studies. In particular it reveals what I call the spatial strategy of each
subject. This is a stance that is adopted relative to various geographical scales, a
seeking out of certain scales of involvement through either sensation or agency,
and an avoidance of other scales. Some people adopt spatial strategies that reach
out beyond the locality, towards global economic, political, and cultural systems.
Others seek to strengthen economic, political and cultural ties to locality. Within
this general range of spatial strategies, from globalizing to localizing, there are a
range of sub-strategies, aligned with various business or professional goals, per-
sonal interests, and life histories, attached to various scales of personal extensibil-
ity.
In the past decade, political geographers have begun to pay increasing attention to
the social construction of scale. Scale is not, on this account, an inherent quality of
interactions, but rather a social product growing out of the 'scale politics of spati-
ality' (Jonas 1994). A labor struggle, for example, is not local, regional, or na-
tional by its nature but, instead, takes on such scale characteristics through proc-
esses of social contestation. Parties to a conf1ict seek either to expand or reduce
the scale of the conflict, depending on their social status. The expansive or con-
tractive tactics of groups can be predicted on the basis of their situation in the con-
flict, such that movements resisting the dominant authorities will often try to
broaden the scale of social involvement, while authorities will strive to constrain
involvement to their territorial jurisdiction (Adams 1996).
Social conflicts make the constructedness of geographical scale abundantly
clear, but more mundane and everyday social processes also manifest the con-
struction of scale. Robert Sack explores this topic in both Human Territariality
220 P.C. Adams
(1986) and Homo Geographicus (1997). In the more recent work, he argues that
the familiar events of daily Iife, such as cooking a meal for guests or helping a
child with school work, provide opportunities for exploration of meaning, nature,
and social relations, and these in turn involve processes and phenomena at a wide
range of scales. Likewise, Jurgen Habermas argues that communicative action
constantly defines and redefines the 'horizon' of the lifeworld, the range of phe-
nomena involved in a particular situation.
An argument, for example, may arise in which one participant justifies an action
as a private choice (bounded) while the other participant criticizes it as a violation
of universal moral principles (unbounded), or one conversant may justity an action
as a response to international social conditions while the other justifies the action
as a response to family problems. So speech constantly reconstructs and problema-
tizes the issue of scale. It is clear that economic relations and political actions do
this as weIl. In fact, all communication does so.
Communication constructs scale in several ways. The two I will explore here are
distant sensation and indirect action. First, sensation is actively extended through
various audio and visual media. I watch a television news feature on the war in
Bosnia or current controversies in the nation's capitol, I read an English novel, or I
listen to a friend in another state who teIls me over the telephone about his new
job. Second, action or agency is extended through institutional frameworks and
media. I order a dozen widgets for my latest professional contract, call my wife to
tell her to buy tomatoes for dinner, or sign a permission slip letting my child go on
a field trip. In these cases, it is virtually assured that agency was guided by sensa-
tion, in the form of viewing apart description in a catalog, exchanging greetings
with my wife, or reading the schedule of the field trip. Less obviously, sensation
implies agency, since sensation is the basis of knowledge, and what we know (or
think we know) affects our actions and the actions of others with whom we com-
municate.
This study applies two different micro-scale lenses to understand personal ex-
tensibility, and more generally, the social construction of scale. Both lenses are
directed towards the lives of five people who live in the Albany, NY metropolitan
area. First, a narrative lens provides a general feel of the subjects' different life-
styles, which range from a quiet retirement to a frenetic schedule juggling several
high-level professional positions. Also included in this narrative are the intersec-
tions of these individuals' life-paths in the experience of one subject who inte-
grates the other four to form a social network. This network is no more important
or real than dozens of others that could be shown involving these persons with
others. It is simply one particular coming-together of people's extended agency
and sensation in an ordinary American city in the 1990s.
CAD-based Accessibility Model 221
Second, models of the subjects' daily routines are shown. These models are in
fact virtual objects constructed in the abstract representational space of a com-
puter-aided design (CAD) program (see Appendix). Since they are stored as a set
of objects in a database, the models can be rotated and examined from various
angles and can be queried to reveal selected themes. For example, one thematic
selection would show only two-way communications such as telephone, and dis-
regard one-way communication links, such as radio and television; altematively,
one and two-way communications could be shown together but in a way that dif-
ferentiates them by appearance. This extensibility diagram is like a GIS database
in that it has no given appearance, but elements can be selected, oriented, and or-
ganized by the user. Furthermore, like GIS, the representation renders a large
amount of data available for analysis rather than condensing the data into statisti-
cal measures, such as mean and standard deviation, which result in a loss of in-
formation.
The goal of the study is to show a way that emerging technologies can be linked
to individual spatial strategies and interpersonal power relations. The sampie is too
small to provide more than a hint of what one might find with a more satisfactory
sampie. The methodology is entirely new, so it is a pilot study. Even so, it hints at
how social power and extensibility are related. The relationship is not a simple
matter of quantity, with more power implying more extensibility; although the
most powerful subjects appear to maintain economic, political, or social power
through extensibility, we cannot reduce that power to a simple metric, such as
more time communicating with distant places or more frequent connections with
distant places. Power is expressed in very different rhythms of extensibility. Fur-
thermore, the less powerful subjects maintain a high level of involvement with
scales of social integration beyond the locality - state, region, nation, and/or
world. Their ability to act at a distance is limited and directed by others, but they
use media to extend their ability to sense the world in ways that fit in with coher-
ent overall spatial strategies.
The powerless are not simply receivers of distant information. They circulate
perceptions in local social contexts that are drawn from distant origins on the
Web, in newspapers, on the radio, and in books. Although minimally capable of
affecting social processes at these sc al es, they are informed about non-Iocal events
and become primary sources of shared local information regarding non-Iocal
events. If this information is distorted by the news sources, as indeed it must be,
people nonetheless are not passive dupes of the media. l They use media to estab-
lish a certain meaningful relationship to the world, and do so in active ways: com-
bining sources, seeking out information related to interests, mixing news and en-
tertainment, comparing news sources, and sometimes regarding the sources they
depend on with a jaded and skeptical eye (Fiske 1987).
1 I am assuming here that all news is a social construction and therefore perpetuates certain
bias es in the way it constructs world events.
222 P.C. Adams
Five subjects were chosen for this study in a non-random way.2 Although too
small a sampie to indicate general patterns, a range of intriguingly different time-
space routines was indicated. This range sufficed to indicate some questions re-
garding the simple assumption of information haves and information have-nots
that has driven other studies of accessibility and IT. In addition, practical aspects
of the representation of extensibility in a CAD-based extensibility diagram could
be explored. Such an approach cannot tell the whole story (even if expanded con-
siderably), but it can help overcome the limitations ofmacro-scale studies.
Diann works in her horne, primarily doing light assembly work for Thomas'
company and caring for her 3-year-old daughter. In her spare time she designs and
creates textile art for sale at a small gallery and invests in commodities. Mr. Wor-
ley, her neighbor, is a retired widower who spends much of his time at horne. He
spends his days reading the newspaper, exercising, taking long walks, and running
errands in his 15-year-old station wagon. Martin Kroopnick is the general manager
of a public radio station with branches throughout New England and upstate New
York, who is also a college professor and the host of a weekly television show.
Thomas owns and manages a small business that designs and produces promo-
tional materials for police departments, primarily a child security kit that helps
parents gather information that will assist police in identifYing a child who is lost.
Lisa is Thomas' secretary; she has a degree in religious studies and is looking for
a job that better uses her skills and contributes more to furthering her social val-
ues. In her spare time she reads esoteric religious texts and does volunteer work
for an environmental organization.
The five people are part of a social network with Diann at the hub. Thomas is
Diann's employer; Lisa is Diann's most common work contact; Martin Kroopnick
is the general manager of Diann's favorite radio station; Mr. Worley and Diann
regularly greet each other when they pass on the sidewalk. The five have obvi-
ously different levels of social power. Martin Kroopnick and Thomas are both, in
different ways, powerful social agents. They both direct the work activities of
dozens of other people and identifY their careers with the achievement of social
power, influence, and success (though the three are mixed in different propor-
tions). They both seem pleased with their career achievements, if also dissatisfied
by some aspects oftheir lives. Diann, Mr. Worley, and Lisa are all near the bottom
of conventional scales of social power. Mr. Worley is marginalized by his age and
his previous employment status, as well as by his comparative lack of education.
Diann and Lisa are professionally marginalized, serving the needs of a business
2 The subjects were all known to the researcher prior to the study and were chosen on the
basis of their willingness to participate and diversity of lifestyles. With a funded study a
larger population could be studied and more reliable results could be obtained. All subject
names have been changed, but are intended to indicate social status in a way that often
occurs in society. Whether a subject is referred to by first name, last name, or both names
mirrors the way the actual subjects were known to the hub subject, Diann.
CAD-based Accessibility Model 223
process that they are not able to guide or direct. To the degree that their work en-
tails communication, that communication is not so much a facet of their own ex-
tensibility as their bosses extensibility. In the terminology ofCastells (1996,244),
these two are the operated rather than the operators, integrators, designers, re-
searchers, or commanders of the information age. These are the executants who
work for the executives. The three 'powerless' subjects differ in that while Diann
and Mr. Worley do not use information technology at all in a professional capac-
ity, Lisa's job is heavily dependent on IT. Here we have what Castells (1996, 244)
calls a 'networked' worker, one who is 'on-line but without deciding when, how,
why, or with whom.'
The five can be contrasted by several aspects of extensibility: (a) the frequency,
duration and overall time devoted to travel, (b) the frequency, duration, and over-
all time devoted to incoming communication (e.g., reading), (c) the frequency,
duration, and overall time devoted to outgoing communication (e.g., writing).
These three general states combine to form a time-space rhythm that is unique to
each individual. In addition, the geographical range of both incoming and outgo-
ing communication varies, as does the social impact of the communications; work-
ing a crossword puzzle and editing a quarterly report are both ways of responding
to incoming communications, but the former clearly has less social impact than
the latter. While it is obvious that the powerful subjects enjoy their power, it is
also evident that the less powerful subjects have qualitatively different spatial
strategies, and this makes it hard to determine who is beUer off in an absolute
sense.
Thursday
weIl as editing a newsletter about the state legislature and teaching three college
courses. Martin's extraordinary schedule suggests an exceptionaIly high level of
extensibility. It is surprising, therefore, that much of his time is spent simply re-
ceiving one-way communications - reading.
tvlart in Kroopnick
12 :0 0 1
mid I I
I I
li sten t 0 radio (own s ta ti on)
12:00
noon
I and repor ts
[
drive to televi sion
I
studi o, record TV
news commen tary ,
drive home
6 :00
pm
I wat eh TV and read
mOOi li ty --"-
one way
12 :00 communicalions { ,...1_ _ _- ,
mid 1 t woway
Figure 13,1. Extensibility diagram indicating the subject's mobility and communications
far a typical Thursday in autumn 1997.
He wakes up to the sound of his radio station and continues to listen while eat-
ing breakfast and driving west across the New York state border to work. When he
arrives at his office he reads the e-mail, which includes communications from ra-
dio station employees to hirn, and communications to station employees from per-
CAD-based Accessibility Model 225
sons outside the station. He forwards the in-coming e-mail and answers the inter-
nal e-mail. Next, he begins to scan a total of seven daily or weekly newspapers
that have arrived at his desk that morning. A break occurs around 9:00 a.m. when
he and his producer drive to a television studio in an adjacent municipality to re-
cord his Sunday television program. He returns an hour and a half later, handles
details of the radio station administration, and continues skimming the papers.
During the ride back to the station he listens to his own radio station, discussing a
program and forming a link to a different space than the one his car moves
through. The metropolitan space shifts to the background of his consciousness and
he attends to the regional space covered by his radio stations.
The afternoon brings a shift of pace with office hours at the university and a
lecture in his journalism class. At 5:30 p.m., when class is over, he drives to the
television studio and prepares for his nightly political commentary. He waits dur-
ing earlier parts of the 6:00 p.m. news broadcast, chatting with the photographer,
the news anchor, and others on the set. He is on the air for two minutes, a span
oddly out of proportion to the familiarity it brings hirn as a political commentator.
He is in his car by 6:30 p.m., heading horne to the small Massachusetts town
where he sleeps.
Martin is in a small minority of the population whose outgoing communications
shape public perception of issues, a role that cannot be measured strictly in quanti-
tative terms, such as in minutes of communication. In essence, he serves a gate-
keeper function (McQuail and Windahl 1981, 100-101; White 1950), influencing
the kinds of issues that will be in the public eye. To be an opinion leader it is not
necessary to hold public attention for long periods of time. What matters are the
channels in which one communicates. Martin reflects this awareness in his obses-
sion with his au dien ce and potential ways to expand it.
Also noteworthy is his mobility. He drives for an hour to his office in Albany
from his house in a small Massachusetts town, rides with his producer for 15 min-
utes to and from his appointment at the television studio where he records his
weekly television pro gram, drives for 10 minutes between his studio and the uni-
versity campus, drives for 15 minutes to return to the television studio to record
his daily news commentary; then drives for an hour back to his house in Massa-
chusetts in the evening. It is not unusual for hirn to spend as much as three hours
on the road during the course of a day.
With such a varied set of responsibilities and ahorne far from his work places,
Martin's time away from horne is elongated beyond the typical range. He leaves
horne around 5:30 a.m. and returns at 7:30 p.m. Martin explains that when he en-
counters members of his audience in the town where he lives they are surprised to
find that he lives there. The time is short in which to enjoy the locational benefits
that his income and power provides. In effect, his activity at the regional scale
necessitates that he detach hirnself from local and metropolitan attachments.
Diann's working hours are even more unusual than Martin's (Figure 13.2). She
begins work at 5:40 a.m. in her living room, assembling child security kits for
Thomas' company while surfing the Web and listening to the radio. Oscillating
between homework and childcare, her working hours extend clear to bedtime, but
226 P.c. Adams
work is interspersed throughout the day with non-work activities, such as walks to
and from her daughter's pre-school, errands, visits with her friend, and 'time out'
to read a novel. The day of the study she takes a bath at 10:40 a.m. and attends a
sing-along with her daughter at 4:00 p.m. More non-work time appears at a finer
scale, in the numerous fluctuations between work and leisure that permeate her
day on a minute-to-minute basis. Even during working hours, her attention con-
stantly shifts between her work, her four-year-old daughter, and the Web or other
media that she uses to occupy her eyes and mind while working on kits with her
hands. The evening of the study day is a bit unusual in that her family runs several
errands in the car: they attend a photography session so they can send personalized
Christmas cards, they visit Thomas' office where Diann drops off the day's pro-
duction, and they shop at the grocery store. Diann lets her husband drive because
she hates driving in the city; she prefers to read a book and ignore the city entirely
until they reach their destination.
Diann's involvement in paid work at the same time she is using telecommunica-
tion and broadcast media for personal pleasure and exploration indicates a rather
unusual spatial strategy. She is not telecommuting, since her use of computer net-
works is for purposes other than work; but, like telecommuters, she is trading ex-
tensibility for mobility. She spends less time than any of the other study subjects
communicating at a face-to-face range, and more time engaged in national and
international scale communications. Her low mobility indicates a personal spatial
strategy of localization, but her use of media such the World Wide Web and Na-
tional Public Radio indicates a strategy of globalization. The horne, traditionally
private, domestic, and local, has become for her somewhat public, professional,
and non-Iocal.
Her social influence is primarily local, but this view of personal scale is prob-
lematized by her extensibility pattern, including the special role she serves as an
information gatherer and community member. She compares the online news with
stories on National Public Radio to get a better sense of non-Iocal events. This
active media use is important to her sense of autonomy and individuality, and adds
validity to her information-gathering role in her immediate social network, which
includes mainly family and friends. Even her boss, Thomas, has consulted her
regarding the Y2K bug, the world economy, and other issues.
Her pleasure with this extensible lifestyle is indicated as she cites Web surfing
as a reason she prefers to work at horne. The choice to stay at horne reduces the
time she spends driving, but that does not reflect a complete withdrawal from her
urban environment. In fact, an unusual proportion of her total travel time is spent
walking, which allows her to attend to her surroundings more than someone who
drives, particularly if they listen to the radio. In addition, she patronizes small
shops in her neighborhood, helping maintain the nodes of the community's activ-
ity. A quantitative comparison of time spent physically moving through the city
would obscure this qualitative difference. Summing up these characteristics we
find a spatial strategy combining attachment to the local - horne and neighborhood
- and the global, with avoidance ofthe metropolitan and state scales of activity.
CAD-based Accessibility Model 227
Diann
12 :00 1
mid
intemet
radio
I intemet, radio
6 :00
am
I walking
reading
television
12 :00 1
noon
I
photo session
shopping
6 :00
pm
1 television
mooi li ty
12 :00
mid
communi cations { 'Ir---. one way
t woway
Figure 13.2. Extensibility diagram indicating the subject's mobility and communications
for a typical Thursday in autumn 1997.
Thomas has a more 'normal' work life than Diann, with a clear phase ofwork-
related extensibility starting shortly after 9:00 a.m. and ending around 5:00 p.m.
(Figure 13.3). This 'window' of activity in time-space defines a great deal of his
interactive (two-way) professional communication. The 9:00 a.m.-to-5:00 p.m.
work period is evident in his routine in the form of frequent, short communica-
tions (mainly by phone) at the national sc ale, altemating with even shorter com-
munications (by voice) at the proximate scale (in his office). He roams his "rG, ..
all day saying: 'Did you mail the order yet?' 'Is this your coffee?' or '!! "i " ~
228 P.C. Adams
Thomas
12 :00 1
mid
6 :00
am I dr i ve and li slen 10 car radi o
1
dn ve and li slen 10 car radio
6 :00
pm web surl
watch television
12:00
mid communicati ons {r
1 _ _ _-,
I
one way
t woway
Figure 13.3. Extensibility diagram indicating the subject's mobility and communications
for a typical Thursday in autumn 1997.
'What's Up?' Work spills over into the rest ofhis day in the form ofletters to read
and accounts to check, either ofwhich can keep hirn busy until 9:30 p.rn.
Two-way communications at the national scale, phone conversations with cli-
ents, dorninate his daily schedule. These are prirnarily routine negotiations with
c1ients involving issues such as the layout of custornized kit covers, production
schedules and fees. While none of his communications are politically powerful in
the way that Martin's are, their existence creates job opportunities for about a
CAD-based Accessibility Model 229
dozen employees and are therefore economically powerful. While Lisa and Tho-
mas may spend equal amounts of time communicating at the national level, Tho-
mas' communications are more closely tied to the exercise of personal power.
Lisa's communications may be better represented as part ofThomas' extensibility.
Thomas is consumed by his work and knows almost nothing about political af-
fairs, environmental problems, or the world economic situation. He knows Martin
Kroopnick only by name. When he hears of a major world event he occasionally
asks Diann or Lisa for their opinion, but otherwise does not search out informa-
tion. Thus, he is active at a large geographical scale, but only in a very limited
way, and with more agency than sensation.
His routine involves a high level of vehicular mobility. In the afternoon he
drives his new Volvo proudly around town doing office errands. While he could
send one of his employees, he values this chance to get away from the office and
enjoys showing off his car and driving aggressively. His car radio is turned up
high and tuned in to a local station with 'oldies' that provides a connection back in
time to his youth. The station draws from a nationally shared list of 1970s hits
thereby also connecting hirn outward in space to anational culture of 30-
somethings. Although Thomas might be conscious of the city of Albany as he
drives, his use of the radio limits the depth to which he is aware of his surround-
ings; he is as much in the music's ambiguous time-space as in the city (a string of
unfortunate car accidents attests to that fact). The telephone at the national scale
constitutes Thomas' professional persona while the car radio at an ambiguous spa-
tio-temporal scale allows hirn to feel at horne.
Thomas' secretary, Lisa, must adhere to a strict schedule, particularly in the
mornings; she is usually the first person to arrive at the office and must be there to
open the office at 9:30 a.m. (Figure 13.4). Still, she wakes up later than all but one
study subject. Around 9:25 a.m. each morning she unlocks the door, turns on the
light, and begins printing a computer report ofthe previous day's sales. While her
duties are varied, the ringing of the phones is the most persistent claim on her at-
tention, followed by the preparation of packages for UPS shipment. Strictly speak-
ing, these activities are part of Lisa's extensibility; she speaks with people in hun-
dreds of police offices around the country and sends them sampies of child-
security products and finished orders. But her actions are even more narrowly de-
fined than are those of Thomas - aside from her personable manner on the phone,
she has little control over what, when, or how she communicates. All of her ac-
tions are determined during work hours by office routines that have developed in
an ad hoc way, or by the logic oftrying to expedite orders, or (less often) by direct
orders from Thomas. Her extensibility during work hours, therefore, is not entirely
her own.
Nevertheless, unlike the automaton stereotype of the information age drone, or
protosurp (Dear and Flusty 1998), Lisa appropriates this situation and retains her
humanity in and through the technologies she must use for her job. She genuinely
enjoys speaking on the telephone to the office's many clients, service providers,
and employees. She brightens up their days with a bit of conversation, and re-
members many people's names and personal trivia in this telephone space.
230 P.c. Adams
Lisa
12 :00 -,-
mid
I
6 :00
am I answe rtelephone,
mail packages. pr i nl
compuler liles,lall<
wilhco-workers
read book dur ing lunch
same as above
I
read book
6 :00
pm
1
I alk 10 housemale, work on puzzle
12:00
mid one way
communicalions { I ___...,
....
1 I woway
Figure 13.4. Extensibility diagram indicating the subject's mobility and communications
for a typical Thursday in autumn 1997.
Outside of business hours and during her fugitive lunch periods (when she
physically leaves the office and goes to the grocery store deli to avoid being inter-
rupted), Lisa enjoys a very different kind of extensibility. She reads broadly from
the texts of many different religions, expanding her knowledge in the subject in
which she holds a Bachelor's degree. At present she is reading the Hermetica and
a book of poetry about trees. She loves books, reading, and talking, and accord-
ingly spends several hours a day engaged in conversation with her boyfriend about
topics of interest from her reading. These emotional and communicational ties to
realities outside the work routine could easily be overlooked as unimportant to the
CAD-based Accessibility Model 231
material fact of her status as secretary, but they provide intellectual sustenance to
remain a friendly voice on the phone as she looks for a job that better uses her
college degree. 3
Her mobility is somewhat limited in speed and distance as she walks to and
from work as weIl as to and from her lunchtime retreat at the grocery store deli.
Qualitative issues are involved. She explains that the daily walks give her time to
think about her poetry and her reading. She sometimes walks past her apartment or
takes a longer route to extend this period of time. For her (like Thomas but in a
slower framework) mobility is obviously more than simply a utilitarian concem
and getting there is not the main objective. However, like Diann, her involvement
with her surroundings seems to be higher that that of Thomas as she moves
through town. Her spatial strategy is interesting.
She maintains strong ties to religious ideas originating in distant times and
places, and to her horne and the section oftown where she walks. While her work-
related extensibility is frequent and economically important, both to her and her
business, it is questionable whether it should even be included in her time-space
diagram or whether it perhaps should be shown in Thomas' diagrarn.
Mr. Worley has the most leisurely schedule (Figure 13.5). He rises any time
between 6:00 a.m. and 7:30 a.m. most days, keeping a rather loose schedule. His
days are filled with a small repertoire of activities: driving to the local Friendly's
restaurant for breakfast, walking in the neighborhood, exercising on the mat in his
TV room, reading the newspaper, watching TV alone or with his adult son, and
buying groceries.
The pace is leisurely, as any one of these activities is drawn out to last at least
half hour. The only activities of short duration are conversations with neighbors,
such as Diann, whom he meets on the sidewalk. Even these can take half an hour
if the other person is not in a hurry. His conversations serve a community-
sustaining function, at least among the older residents of the neighborhood who
have known hirn for years. In this sense, he manifests a localizing spatial strategy.
Nevertheless, Mr. Worley is not out of touch with the world. He spends several
ho urs a day reading the newspaper and informs hirnself of key events in his local
community and a smattering of national affairs that catch his attention. Most of his
communication is one-way communication: reading the newspaper or watching
television. His two-way communication opportunities are primarily in face-to-face
situations, with neighbors, Friendly's employees, and his house cleaner. Unlike
Diann, he does not act as an information source. Some of what he learns he keeps
to hirns elf. More often, he takes delight in bringing his discussions around to his
cynical and ironic perception that the world is in a mess.
Less deliberately than Diann, he has traded mobility for extensibility. His
housebound lifestyle increases his sensory involvement in affairs at the metropoli-
tan, regional, and national scales. Attitudinally, however, he is not particularly
3 By the time of this writing, she has found another job organizing the planting of 'cham-
pion' trees, which conforrns more closely to her values and goals. Presumably her reading
and writing were preludes to this career path.
232 P.C. Adams
receptive to such distant information, often summing up his view of the world
with a few favorite expletives. This, too, is a spatial strategy. It is a brand of local-
ism that draws parasitically on non-Ioeal media as a source of ironie eritique of
distant events.
Mr. Worley
12 :00 1
mid
I
read newspaper
6 :00
am
I" shop
r ead ne wspape r
r ead A AFf> bu II el in
and read mai I
12 :00 1
noon
I walch
televi sIon
6 :00
pm
1 ta lk 10 son
walch
televi sion
12 :00
mid communi calions { I,. .___• one way
I I woway
Figure 13.5. Extensibility diagram indicating the subject's mobility and communications
for a typical Thursday in autumn 1997.
CAD-based Accessibility Model 233
13.4 Discussion
These five people have greatly divergent patterns of extensibility. Their differ-
ences are only superficially tied to opportunity. Strictly speaking, Mr. Worley and
Diann have similar opportunities to communicate but Diann spends much of her
time listening to the radio and surfing the Web while Mr. Worley spends his time
reading newspapers and chatting with neighbors. His lack of a computer is due not
to a lack of funds or proper location, but rather a lack of interest. Nor can the dif-
ference be reduced to a difference in cultural capital without recognizing that Di-
ann is not profiting from her use of the Internet but simply using it to pass the time
and to situate herself in the world. Mr. Worley simply satisfies these objectives in
a different way. To construe her greater degree of connectivity as an advantage is
to misconstrue her purpose and rewards, and to impose an apriori judgment on
these two forms of communication. In contrast, sometimes-similar levels of exten-
sibility are not what they appear to be. Lisa and Thomas have much the same
communication opportunities during work hours, but one determines what will be
communicated while (with certain caveats) the other does not. Surely Thomas has
an advantage in extensibility, although that advantage is difficult to specify.
What has been called accessibility is most often based on measures of opportu-
nity derived from technological and economic patterns, but this is clearly an in-
adequate notion. Communications more closely reflect social power, but here
again the connection is subtle. Most important are personal interests, goals, habits,
and social connections, phase in the life cycle, gender, and other factors of the
individual.
Communication is, in short, part of a time-space routine that is personal and
difficult to generalize. The idea of 'opportunity' as typically construed in studies
of accessibility must, therefore, be supplemented by actuality. Mr. Worley gets aB
the newspapers he needs, and what he misses on the Internet generally lies beyond
his range of interests. Thomas' business depends on the telephone, but aside from
purchasing several independent lines and a fax machine he has not yet feIt a need
to adopt more sophisticated telecommunications at work. He does have Internet
access from his horne, but uses it infrequently. Lisa's passion is communication,
but not telecommunication; she prefers a good book that she can carry out of the
office during her lunch hour, and discuss in the evening with her boyfriend as they
sit in the kitchen or bedroom. Martin uses the Internet daily, but mainly for receiv-
ing e-mail; he keeps abreast of current events by reading national, regional, state,
and local newspapers. Diann alone has built the Internet into her life as a serious
passion, not because of her social status, but in part through the acceptance of a
job with little social status.
Likewise, the most mobile subjects, Martin and Thomas, do not appear to be in
the space they are so often driving through. They listen to the car radio, one to
survey his own indirect labor, the other to bind past and present, dream and actual-
ity. Physical mobility is, oddly, a constraint they both seek to overcome with ex-
234 P.C. Adams
tensibility. I would argue that what is overcome is the involvement with physical
surroundings that is possible when one walks through an environment, like the
apparently less mobile subjects. Again, qualitative issues complicate the picture:
access to space is not merely measurable in terms of distance; one must also con-
sider depth.
A map of communication opportunities, for example a map of Internet hosts
and data transmission backbone, misses an essential point. Opportunities are
determined not by an abstract calculus but by individually determined needs.
Constraints are constraints only when they interfere with someone's self-defined
needs. Only theoretical individuals can occupy the spaces of opportunity shown
by traditional maps of accessibility. The five individuals in this study display
five radically different lifestyles and five ranges of need, these correspond to five
spatial strategies of real persons. Older media such as books and telephones eas-
ily satisfy some oftheir spatial strategies, while others require newer media, such
as television and the Internet. The abstract metric of accessibility cannot capture
the relativity of individual needs and goals.
It is notable that the connections between Diann and the other four subjects are
brief. If one aggregates all such interactions, calls them interactions in place, and
compares them with interactions at a distance, the two are roughly equal in time
use. People who spend more of their time at horne spend a higher proportion of
total time involved in non-local communication situations. This points to a trend
of localization within globalization, or perhaps globalization within localization.
The opposition between global and local cannot describe the individuals in the
study because they are folding the world into the space ofthe locality.
For some of the subjects, horne has dwindled in functional importance because
of a mobile lifestyle and heavy workload. In this category are Martin Kroopnick
and Thomas. Perhaps, also, locality has dwindled for these people. While Thomas
lives 15 minutes from his place of work, Martin lives an hour away. His lifestyle
choice is not uncommon among the wealthy. The irony of this choice is that its
costs, the personal loss of 16 hours per week and the public cost of automobile
emissions, are spent on a choice that he does not appear to have much time or en-
ergy to enjoy. Thomas also suffers high costs in the form of the period of time he
is away from horne each day. Diann reveals an alternative strategy that does not
provide the same benefits in social power, but satisfies goals relating to quality of
life and the quality of others' (her child's) life. No doubt women are more inclined
to emphasize these values. Even so, Ü is not necessary that they sacrifice career
goals while adopting this spatial strategy (Helgeson 1998).
Surveying the different rotations of the extensibility diagram (Figures 13.6 and
13.7), we see a complex architecture of ties to the regional, national, and global
scales. This architecture clings tenuously together at the local level, with momen-
tary exchanges such as the 'hello' one says to a neighbor or the brief exchange of
work-related information. The most intimate relationships are not shown, such as
ties between husband and wife, best friends, and associates who work in the same
office. Nevertheless the diagram suggests a truth about modem urban settings.
CAD-based Accessibility Model 235
Almost any five citizens drawn randomly from the Albany metropolitan area will
be held together tenuously if at all by communication links.
Diann says "Good
morning" 10 Mr. Worley.
Mr. Worley
Usa
Diann haars Mari in
Kroopnick on I he
radio.
Diann
Figure 13.6. Extensibility diagram combining Figures 13.1 - 13.5, and also showing com-
munications between Diann and other participants
In the information age, personal ties to non-Iocal scales seriously riyal ties to
locality. What makes this situation perplexing is not its existence, because that
could be attributed simply to the scale ofthe extant communication media. Rather,
its mystery lies in the way people mix radically different spatial strategies in the
pursuit of personal goals and seem to find mutually agreeable results in the proc-
ess. There is a process of specialization that is entirely different than the speciali-
zation of skills employed in one's job: people also specialize in the spatial strate-
gies they employ in the pursuit of personal and collective goals. In a way
analogous to the merging of specialized skills to produce a diversified economy,
we can observe a merging of spatial strategies to produce a diversified space-time
fabric. lt may be that the regional focus of one facilitates the local focus or global
focus of another.
Ties at the macro-scale between accessibility and economic power must be ac-
knowledged. The wealthy and powerful do have more communication tools at
their disposal. However, we must not, as a consequence of this observation, over-
236 P.C. Adams
Diann
Mr.
",
Figure 13.7. Extensibility diagram combining Figures 13 .1-13.5, and also showing com-
munications between Diann and other participants.
GI 100%
.~ 90%
CI 80 %
c
-
:i: 70%
CI!
~ 60%
0 50%
GI 40 %
CI
~ 30%
c 20%
GI
I:! 10%
GI
Q. 0%
« III « III .><
c:
c:
c:
c: ~~
>.
..: Q) .-c ·~
c:
1: a.
'"
Ci
CI!
(5
~"t:
0
~"t:
'" 0
~ ~ ~e
~
Figure 13.8. Approximate pereentage ofwaking hours spent by eaeh partieipant eommuni-
eating at a given range, ineoming and outgoing eommunieations - eombined. Reprinted
with permission from Urban Geography, Vol. 20, No. 4, pp. 356-76. © V.H. Winston &
Son, Ine., 360 South Oeean Boulevard, Palm Beaeh FL 33480. All rights reserved.
References
Adams, P. 1995. A reeonsideration of personal boundaries in spaee-time. Annals 0/ the
Association 0/American Geographers 85(2):267-85 .
Adams, P. 1996. Protest and the seale polities ofteleeommunieations. Political Geography
15( 5):419-41.
Adams, P 1999. Bringing globalization horne: A homeworker in the information age. Ur-
ban Geography 20:356-76.
Castells, M. 1996. The Rise ofthe Network Soeiety. The Information Age: Eeonomy, Soci-
ety and Cuiture, I. Maiden MA and Oxford: Blaekwell.
Castells, M. 1997. The Power of Identity. The Information Age: Eeonomy, Soeiety and
CuIture, 2. Maiden MA and Oxford: B1aekwell.
Castells, M. 1998. End ofMillenium. The Information Age: Eeonomy, Soeiety and Culture,
3. Maiden MA and Oxford: Blaekwell.
Curry, M. 1995. Geographie Information Systems and the inevitability of ethiea1 ineonsis-
teney. In Piekies, J. (ed.) Ground Truth., New York and London: The Guilford Press, 68-
87.
Dear, M., and Flusty, S. 1998. Postmodern urbanism. Annals 0/ the Association 0/ Ameri-
can Geographers 88:50-72.
Fiske, J. 1987. Television Culture. London and New York: Methuen.
CAD-based Accessibility Model 239
Graham, S. and Marvin, S. 1996. Telecommunications and the City: Electronic Spaces.
Urban Places. London and New York: Routledge.
Habermas, 1. 1987. The Theory of Communicative Action. Lifeworld and System: A Cri-
tique of Functionalist Reason, 2. Translated by Thomas McCarthy. Boston: Beacon
Press.
Hägerstrand, T. 1970. What about people in regional science? Papers 0/ the Regional Sci-
ence Association 24:7-21.
Helgeson, S. 1998. Everyday Revolutionaries: Working Women and the Transformation of
American life. New York: Doubleday.
Janelle, D. 1973. Measuring human extensibility in a shrinking world. The Journal o/Ge-
ography 72(5):8-15.
Jonas, A.E.G. 1994. Editorial: The scale politics of spatiality. Environment and Planning
D: Society and Space. 12(3):257-64.
McQuail, D., and Windahl, S. 1981. Communication Models. London and New York:
Longman.
Sack, R. 1986. Human Territoriality: Its Theory and History. Cambridge: Cambridge Uni-
versity Press.
Sack, R. 1997. Homo Geographicus: A Framework for Action, Awareness, and Moral Con-
cern. Baltimore and London: The John Hopkins University Press.
White, D.M. 1950. The 'Gatekeepers' : A case study in the selection of news. Journalism
Quarterly 27:383-90.
14 Human Extensibility and Individual Hybrid-
accessibility in Space-time: A Multi-scale
Representation Using GIS
Mei-Po Kwan
14.1 Introduction
With the increasing use of the Internet for getting information, transacting busi-
ness and interacting with people, a wide range of activities in everyday life can
now be undertaken in cyberspace. As traditional models of accessibility are based
on physical notions of distance and proximity, they are inadequate for conceptual-
izing or analyzing individual accessibility in the physical world and cyberspace
(hereafter referred to as hybrid-accessibility). To address the need for new models
of space and time that enable us to represent individual accessibility in the infor-
mation age, there are at least three major research areas: (a) the conceptual and/or
behavioral foundation of individual accessibility; (b) appropriate methods for rep-
resenting accessibility; and (c) feasible operational measures for evaluating indi-
vidual accessibility. With the recent development and application of GIS methods
in the study of accessibility in the physical world (e.g., Forer 1998, Hanson, Ko-
miniak, and Carlin 1997, Huisman and Porer 1998, Kwan 1998, 1999a, 1999b,
MilIer 1991,1999, Scott 1999, Talen 1997, Talen and Anselin 1998), it is apparent
that GIS have considerable potential in each of these research areas. As shown in
some of these studies, a focus on the individual enabled by GIS methods also re-
veals the spatial-temporal complexity in individual activity patterns and
accessibility through 3D visualization or computational procedures.
Yet, even with the advent of 3D GIS tools, there are several difficulties when
GIS methods are applied to represent or measure individual hybrid-accessibility.
First, personal accessibility in the age of information involves multiple spatial and
temporal scales (Hodge 1997), whereas current GIS are designed to handle only
one geographical and/or temporal scale at a time. For instance, personal extensi-
bility enabled by telecommunication technologies now allows an individual to
access information resources at the global scale although the person's physical
activities are still largely confined at the local scale. Further, the traditional tempo-
ral scale (ho ur/minute ) is not adequate for studying cyber-transactions that may be
accomplished within a few seconds. Second, GIS-based representational and com-
putational methods, such as the space-time prism, are based on the sequential un-
242 M-P Kwan
folding of a person's activities in the physical world. They are not developed to
handle the simultaneity and temporal disjuncture that characterize many types of
cyber-transactions. For example, a person may be talking over the phone and
browsing a Web page at the same time. An email message sent out now may be
read several hours later on the other side of the globe. These limitations of current
GIS methods constitute a major challenge to any effort to represent and measure
individual hybrid-accessibility in the information age.
As a preliminary attempt to address this methodological challenge, this paper
explores how current GIS, given their limitations, can be deployed for the 3D in-
teractive visualization of human extensibility in space-time. It develops and pre-
sents a method for the multi-scale, 3D representation of individual space-time
paths based upon the concept of human extensibility (Janelle 1973, Adams 1995).
Using geo-referenced activity diary data for an individual as an example and Arc-
View GIS software (© ESRI, Inc), the method is capable of revealing the spatial
scope and temporal rhythms of a person's extensibility in cyberspace. It can also
represent the complex interaction patterns among individuals in cyberspace using
multiple and branching space-time paths within a GIS. Compared with the two-
dimensional and/or cartographic representations in past studies, this method al-
lows the researeher to interact, explore and manipulate the 3D scene (e.g., rota-
tion, fly-through). This visualization environment not only greatly facilitates ex-
ploratory data analysis, but can also enhance our understanding of the patterns
portrayed. It may provide the basis for formulating operational measures of indi-
vidual hybrid-accessibility. In this paper, the nature of accessibility in the informa-
tion age is first examined, and then alternative representational methods are dis-
cussed. Implementation of the GIS method using real activity diary data of an
individual is described.
Janelle (1973) first formulated the concept ofthe individual as an extensible agent,
where extensibility represents the ability of a person to overcome the friction of
distance through space-adjusting technologies, such as transportation and commu-
nication. As the conceptual reciprocal of time-space convergence, which reflects
the degree to which places are approaching one another in time-distance, human
extensibility measures the increased opportunities for interaction among people
and places (Janelle 1973). The development of communication and transportation
technologies (or spatial technologies) and their associated institutions thus imply a
shrinking warld with expanding opportunities for extensibility (Adams 1995,
Coucleclis 1994). Further, human extensibility not only expands a person's scope
of sensory access and knowledge acquisition, it also enables a person to engage in
distantiated social actions whose effect may extend across disparate geographical
regions or historical episodes (Adams 1999, Thrift 1985).
Adams (1995) extended this notion of human extensibility through a new model
of the person based on the structuration perspective (Giddens 1984), where the
spatially contingent and socially embedded nature of human extensibility is em-
phasized. Inequality in human extensibility with respect to gender, race and other
socially significant categories is understood in terms of the mutually constitutive
relations between the individual experience of accessibility and macro-level socie-
tal processes. Adams (1995) captured the dynamic and fluid nature of personal
boundaries through the notion of 'people as amoebas'. The body is reconceptual-
ized as a dynamic entity, which combines
a body rooted in a particular place at any given time, bounded in knowledge gathering
by the range of unaided sensory perception, [and] ... any number of fluctuating, den-
dritic, extensions which actively engage with social and natural phenomena, at varying
distances (Adams 1995, 269).
This notion of human extensibility not only provides a useful point of departure
for understanding individual accessibility in the information age. It also offers a
theoretical foundation for overcoming many limitations in the traditional under-
standing of corporeality found in Hägerstrand's time-geographic framework. As
Rose's (1993) critique suggests, depicting a person's trajectory in space-time as a
linear and clear-cut path has many difficulties, especially when the framework is
used to understand women's everyday lives. Further, since constructs of the time-
geographic framework have been used to formulate accessibility measures in the
past (e.g., Bums 1979, Lenntorp 1976, Villoria 1989), a representational device
capable of handling this reconceptualized extensibility is an important first step in
formulating operational measures of individual hybrid-accessibility.
For this purpose, Adams (1995) developed the extensibility diagram using the
cartographic medium. The diagram, based on Hägerstrand's space-time aquarium,
portrays a person's daily activities and interactions with others as multiple and
Human Extensibility and Individual Hybrid-accessibility 245
after work. These kinds of interactions between activities in the physical world
and cyberspace would be difficult to identifY when this method is used.
Because of the !imitations of these two methods, another method for the multi-
scale, three-dimensional representation of individual space-time paths in hybrid
physical-virtual world is discussed in the next section. This method integrates
transactions at different spatial scales in one graphical window, where the overall
pattern or relationships among activities at different spatial scales can be easily
identified. It can represent various types of transactions that take placc at the same
time (simultaneity). Further, the method allows differentiation of attributes for
each transaction using graphical legends. For example, color codes can show
transactions at different spatial scales (Ioeal, regional, global), with different tem-
poral characteristics (synchronous, asynchronous), and undertaken through differ-
ent communication modes (one-way incoming, one-way outgoing, two-way).
Most of the data pre-processing was performed using ARCINFO while the visu-
alization was implemented using ArcView 3D Analyst.
(a) Data. The activity data of an individual, Pui-Fun (a fictitious name), was col-
lected and used to implement the GIS method. This person is a software engineer
who works in a telecommunications company in Columbus, Ohio. Information
about her activities in the physical world was collected in the form of an activity-
travel diary. Data about her activities in cyberspace were compiled from the his-
tory file of her Web browser and email directory. As these data did not come with
time stamps for computing the timing and duration of her cyber-transactions, the
temporal information needed for constructing the space-time path was recon-
structed through a personal interview, in which she also explained each of her ac-
tivities recorded on the diary day (Table 14.1). This makes a GIS-based graphic-
narrative ofher activities on this day possible. Further, as several Web pages were
browsed during each ofher visits to the Web sites recorded, Web browsing activi-
ties are grouped into distinctive sessions identified by the site visited (instead of
presenting details of each page browsed). Table 14.1 divides her activities in terms
of the local, regional (15 northeastern states in the United States) and global scales
according to the location ofthese transactions. The following account ofPui-Fun's
cyber-transactions focuses mainly on the Internet since other forms of personal
extensibility such as interactions via the telephone were not recorded.
248 M-PKwan
Source: Activity diary and personal interview with subject, August 1998
Activity data provide information about Pui-Fun's activities in both the physical
world and in cyberspace for a Saturday that she worked from 8:30 a.m. to 12 mid-
night (Table 14.1). This is an unusual schedule since she normally works only
from 9 am to 5:30 p.m. on weekdays. On this day, Pui-Fun's husband dropped her
off at work at about 8:30 a.m. As she was working with several co-workers in
other branches of the company located in Chicago IL, Maywood NJ and Charlotte
NC to meet the delivery deadline of a product, she regularly checked the company
Web pages on which important news and updates about the project were posted.
Further, she exchanged several email messages with these co-workers throughout
the day since there might be last-minute debugging and testing tricks she needed
to know for preparing the final shipment of the product to the client. On the diary
day, Pui-Fun started her day at 8:35 a.m. with a brief session of Web browsing at
the Chicago site. Then, shortly after, she browsed more extensively to make sure
she did not miss anything important, covering the Chicago, Maywood, and Char-
lotte sites. After she got all the necessary information, she continued to work on
the project. She only had a brieflunch break at her workplace at about 12:30 p.m.
Around 2:00 p.m., she came across a technical problem which required her to
log onto the project's information site in Chicago again. In late aftemoon, she
conducted another round of routine browsing of the Charlotte, Chicago, and
Maywood sites. Because time is so limited for meeting the project's deadline, Pui-
Fun did not go out for dinner on this day. Instead, her husband brought her dinner
from a fast-food chain at about 6:30 p.m. She stayed at her workplace for the
whole day and was off at about 12:00 midnight. Because company policy restricts
Human Extensibility and Individual Hybrid-accessibility 249
employees' private use of the Internet at work, she rarely reads personal email
messages or browses her favorite Web pages during work hours. She usually does
so before her formal work hours begin or in the twenty minutes or so after she is
formally off from work while waiting to be picked up by her husband. On this
day, she browsed so me newspaper and magazine Web pages hosted in Hong Kong
while her husband was on the way to pick her up.
Figure 14.1. A two-dimensional representation ofthe three map layers after transformation.
Hang Kong
== == == == == =: i
Charlotte Maywood
Chicago
e paths.
esentation ofth e individual's space-tim
Figure 14.2. A multi-scale, 3D repr
252 M-PKw an
9 --
--..-:::.,. .........
-- ............ ....
...... , .... . . .
--
.
Figure 14.3. An extensibility diagram of a set of hypothetical activities
the po-
Given the limited range of Pui-Fun's activities on the actual diary day,
so me hy-
tential of this GIS-based extensibility diagram is further explored using
transact ions at
pothetical activities. The objective is to show how various types of
as an exampl e and
different spatial scales can be represented. Using Pui-Fun
Human Extensibility and Individual Hybrid-accessibility 253
partly following Adams' (1995) scheme, Figure 14.3 shows five types ofactivities
undertaken on a particular day. On this day, Pui-Fun worked from 8:30 a.m. to
5:30 p.m., and had a one-hour lunch break at a nearby restaurant (c on the dia-
gram). She subscribes to a Web-casting service where news items are continu-
ously forwarded to her Web browser. On this day, she read some news about
Yugoslavia, South Africa, and Nashville, TN (a on the diagram) before she started
work. An hour later she sent an email message to three friends located in Hong
Kong, Chicago, and Vancouver (b). The friend in Chicago read the email two
hours later and the friend in Vancouver read the email five hours later. The friend
in Hong Kong read the email 13 hours later and replied immediately (e). The reply
message from this friend, however, was read at 2:00 a.m. at Pui-Fun's horne (g).
In the afternoon, Pui-Fun browsed Web pages hosted in New York, Charlotte, and
Anchorage in Alaska (d). She was off from work at 5:30 p.m. and spent the eve-
ning at horne. At 9:00 p.m. she started an ICQ (real-time chat) session with friends
in Tokyo, Melbourne, Memphis TN, and Dublin OH (f on the diagram).
As shown in Figure 14.3, very complex interaction patterns in cyberspace can be
represented using multiple and branching space-time paths. These inc1ude tempo-
rally coincidental (real-time chat) and temporally non-coincidental (e-mailing)
interactions; one-way radial (Web browsing), two-way dyadic or radial (e-
mailing), and multi-way (chat) interactions; in-coming (Web casting) and out-
going (e-mailing) transactions (Adams 1998, Janelle 1995). The method is thus
capable of capturing the spatial, temporal, and morphological complexities of a
person's extensibility in cyberspace.
There are other difficulties in implementing the method. First, since detailed
data of an individual's activities in physical and cyberspace space are needed for
constructing the 3D extensibility diagram, data availability will be a major issue.
The problem is especially serious for transactions in cyberspace, as there are not
only many different types of transactions to be recorded (e.g., e-mailing, Web
browsing, Web casting, real-time chat, etc.), there is also no readily available
means for recording these transactions. Data collected by commercially available
server-side logging programs (used frequently by computer network administra-
tors) are not adequate for this kind of study. Future research needs to investigate
how to record these activities on the client side. This would involve a major diffi-
culty regarding personal privacy: Will individuals be willing to disclose this kind
of personal information in such detail?
Second, even when data about cyber-transactions are available, the location of a
particular host on the Internet may be difficult to identify since IP addresses may
not map onto geographical locations uniquely. Lastly, although individual space-
time paths can be represented using this 3D GIS method, it renders the computa-
ti on of space-time accessibility measures much more difficult. Given that cyber-
transactions involve multiple spatial and temporal scales, and may include multi-
ple and branching space-time paths, how can the space-time prism be identified?
When fixed activities (such as work) may be ongoing with other flexible activities,
which may involve far-away locations, how should space-time accessibility meas-
ures be computed? Each of these areas requires further research.
Acknowledgements
The support of an NCGIA Varenius seed grant for this research is gratefully acknowledged.
I also thank the person who provided the aetivity data for this study.
References
Adams, P.C. 1995. A reconsideration of personal boundaries in space-time. Annals o/the
Assoeiation 0/ Ameriean Geographers 85(2):267-85.
Adams, P.C. 1998. Network topologies and virtual place. Annals 0/ the Assoeiation 0/
Ameriean Geographers, 88(1 ):88-1 06.
Bums, L. 1979. Transportation, Temporal, and Spatial Components of Aceessibility. Lex-
ington, MA: Lexington Books.
Carlstein, T. 1982. Time Resourees, Society and Eeology: On the Capacity for Human
Interaction in Space and Time in Preindustrial Society. Lund Studies in Geography, Se-
ries B, Human Geography NO.49. Lund, Sweden: CWK Gleerup.
Couclelis, H. 1994. Spatial technologies. Environment and Planning B 21: 142-43.
Forer, P. 1998. Geometrie approaches to the nexus of time, space, and microproeess: im-
plementing a practieal model for mundane socio-spatial systems. In Egenhofer, M.J. and
Golledge, R.G. (eds.) Spatial and Temporal Reasoning in Geographie Information Sys-
tems. Oxford: Oxford University Press, 171-90.
Giddens, A. 1984. The Constitution of Socity: Outline of the Theory of Structuration. Ber-
keley, CA: University ofCalifomia Press.
Human Extensibility and Individual Hybrid-accessibility 255
Hägerstrand, T. 1970. What about people in regional science? Papers ofthe Regional Sei-
ence Association, 24:7-21.
Hanson, S., Kominiak, T. and Carlin, S. 1997. Assessing the impact of location on
women's labor market outcomes: A methodological exploration. Geographical Analysis
29:281-97.
Hodge, D. 1997. Accessibility-related issues. Journal ofTransport Geography 5(1):33-4.
Huisman, O. and Forer, P. 1998. Towards a geometrie framework for modelling space-time
opportunities and interaction potential. Paper presented at the International Geographical
Union, Commission on ModeIIing Geographical Systems Meeting (IGU-CMGS), 28-29
August, Lisbon, Portugal.
JaneIIe, D. 1973. Measuring human extensibility in a shrinking world. Journal ofGeogra-
phy 72(5):8-15.
JaneIIe, D. 1995. Metropolitan expansion, telecommuting, and transportation. In Hanson, S.
(ed.) The Geography ofUrban Transportation, ]"'1 edition. New York: Guilford, 407-34.
Kwan, M-P. 1998. Space-time and integral measures of individual accessibility: A com-
parative analysis using a point-based framework. Geographical Analysis, 30(3):191-216.
Kwan, M-P. 1999a. Gender and individual access to urban opportunities: A study using
space-time measures. The Professional Geographer 51(2):210-27.
Kwan, M-P. 1999b. Gender, the home-work link, and spaee-time patterns of non-
employment activities. Economic Geography 75(4).
Kwan, M.P. 2000. Cyberspatial eognition and individual aceess to information: The behav-
ioral foundation of cyber-geography. Environment and Planning B, in press.
Kwan, M-P. and Hong, X-D. 1998. Network-based constraints-oriented choice set forma-
tion using GIS. Geographical Systems, 5:139-62.
Lenntorp, B. 1976. Paths in Time-Space Environments: A Time Geographie Study of
Movement Possibilities of Individuals. Lund Studies in Geography, Series B, Human Ge-
ography No.44. Lund, Sweden: CWK Gleerup.
MathSoft. 1997 S-Plus User's Guide. Seattle, WA: MathSoft, Inc.
MCI, 1998. Internet Trafik Report <http://traffic.meLcom/>
MIDS, 1998. The MIDS Internet Weather Report (IWR) <http://www.mids.org/weather/
us/htmldir/>
MilIer, H.J. 1991. ModeIIing aceessibility using spaee-time prism concepts within geo-
graphie information systems. International Journal ofGeographical Information Systems
5(3):287-301.
Miller H.J. 1999. Measuring spaee-time aeeessibility benefits within transportation net-
works: Basic theory and eomputational proeedures. Geographical Analysis 31(2):187-
212.
Pred, A. and Palm, R. 1978. The status of Ameriean women: A time-geographie view. In
Lanegran, D.A. and Palm R. (eds.). An Invitation to Geography. New York: MeGraw-
HiII,99-109.
Rose, G. 1993. Feminism and Geography: The Limits ofGeographical Knowledge. Min-
neapolis, MN: University ofMinnesota Press.
Seott, L.M. 1999. Abstract - Evaluating intra-metropolitan aeeessibility in the information
age: Operational issues, objeetives, and implementation. In JaneIIe, D.G. and Hodge,
D.C. (eds.) Measuring and Representing Accessibility in the Information Age, Research
Conference Report. Santa Barbara CA: NCGIA, 30-31.
256 M-P Kwan
Talen, E. 1997. The soda! equity of urban service distribution: An exploration of park ac-
cess in Pueblo, Colorado, and Macon, Georgia. Urban Geography, 18(6):521-41.
Talen, E. and AnseIin L. 1998. Assessing spatial equity: An evaluation of measures of ac-
cessibility to public playgrounds. Environment and Planning A, 30:595-613.
Thrift, N. 1985. Flies and germs: a geography of knowledge. In Gregory, D. and Urry, J.
(eds) Social Relations and Spatial Structures. New York: St. Martin' s Press, 366-403.
Villoria, O.G. 1989. An Operational Measure of Individual Accessibility for Use in the
Study of Travel-Activity Patterns. Ph.D. dissertation, Graduate School of the Ohio State
University.
Part 111
Societal Issues
15 Accessibility and Societal Issues
in the Information Age
Mark I. Wilson
Department of GeographylUrban and Regional Planning, and Institute for Public Policy and
Social Research, Michigan State University, East Lansing MI 48824-1111, USA.
Email: wilsonmm@pilot.msu.edu
15.1 Introduction
Second, information technologies and their services are not spatially confined, but
are a global medium that appears to be aspatial by avoiding identity with a place
or country. Third, as a new phenomenon, there is a lack of consensus about what
to analyze and measure about the new technologies. Finally, much of the analysis
and measurement of IT services is by individuals, firms, and agencies with a
vested interest in the results, raising issues about standards and impartiality. Given
these conditions, it is important for scholars to explore ways to understand, meas-
ure, and inform others about the nature ofthese new media.
In many ways the technical challenges are waning compared to the social chal-
lenges of access for the information society. World maps may showalmost uni-
versal access, but there is a significant difference between access being possible
and people being able to access the Internet if and when they wish. The non-
technical challenges are many if access is considered integral to membership in
the information society. Social factors that are important include the relevance of
access to many whose daily life may not yet require access; and, education and
experience in using information technologies. Economic factors, such as afforda-
bility; use in employment; and the growing importance of information-based
economies shape access to information technologies. Political factors also 100m
large, with individuals having the freedom to access information; with regulatory
environments supporting access and affordability of access; and government gen-
erating, managing, and using information effectively.
While not a unique element in this volume, Part III focuses specifically on the
social context in which information technologies are used and developed. Implicit
in this analysis is an awareness of the need to understand how society and tech-
nology meet and interact. In fact, the political, economic, and social dimensions of
information technology use may weil represent a far more complex and controver-
sial arena for IT development than the technical barriers that scientists confront
when advancing information technologies. As William Mitchell (1995, 5) notes in
City ofBits,
... the most crucial task before us is not one of putting in place the digital plumbing of
broadband communications links and associated electronic appliances (which we will
certainly get anyway), nor even of producing electronically deliverable 'content,' but
rather one of imagining and creating digitally mediated environments for the kinds of
lives that we will want to lead and the sorts of communities that we will want to have.
valuable to note some of the key questions that need to be considered when assess-
ing the broader use of information technologies.
In the context of this volume, the core questions and issues revolve around ac-
cess to information technologies. At the individual level, this type of access con-
cerns (1) the availability of a nearby computer and the physical ability to use in-
formation technologies by knowing how to use a computer and to access the
Internet; (2) having the resources to buy or rent a computer and to afford Internet
access; and (3) having the freedom to interact with others electronically, or to
view material of interest. As information technology evolves, so me of these con-
cerns will be accommodated. However, there will always be a need to focus on the
broader implications of information technology to social change. The chapters in
Part II of this book highlight five general constructs that weigh heavily in any con-
sideration of the societal implications of IT - social context; equity; rights; time;
and processes.
262 M.1. Wilson
Social Context
This does not mean that the conventional notions of accessibility no longer apply. It
simply reflects the fact that accessibility is an intrinsically manifold notion, encom-
passing several definitions that can co-exist and not be reducible to each other.
Equity
Equity considerations are increasingly important as the ability to gain access im-
proves and the cost of access decreases. If access, control, and management of
information are the foundations of economic growth and development, then the
core equity considerations become: Who has access? And is access possible for
those who desire access? While fundamentally an economic issue of affordability,
Accessibility and Societal Issues 263
the ability to access is also determined by government actions to shape the policies
of service providers, or to make access possible through schools, libraries, and
other public facilities. Emerging from equity concerns are research issues sur-
rounding the relationship between access to information technologies, informa-
tion, and economic and political power. Also relevant is an understanding of how
inequalities change over time, and the ways in which IT affects the rate and direc-
tion of change in an information economy.
Beyond the affordability issues of access lie a number of geographic elements,
as access has long been seen as a question of physical proximity. On one level,
proximity to Internet service remains a crucial factor. Can I access the Internet
where I live or work and at what cost? At a broader level, however, Internet access
can reduce the physical barriers that have prevailed in the past. Information tech-
nologies may be able to remove the barriers that have defined peripherallocations
to date. For example, Ireland's peripheral geography has been overcome in many
ways through public and private investment in infrastructure, education, and train-
ing. The shift from peripheral geography to electronic centrality carries great
benefits to the people and countries that are able to engineer relocation to elec-
tronic space.
Susan Hanson explores how spatial technologies affect equity in Chapter 16. In
particular, she goes beyond issues of physical access to information technologies
to address the social importance of information flows and their form. She calls for
an understanding of how new technologies intersect with existing social relations
in building and maintaining social equity, and she raises important questions about
the social value of information technologies in comparison to face-to-face rela-
tionships and communication through existing community networks.
Rights
The discussion of rights recognizes a broader domain for the importance of free-
dom and the role of ethics in electronic interaction and access. Issues of right to
access and use oftechnology and information are based in political and social con-
text. The choices that societies make in relationship to utilitarian or communi-
tarian systems raise several categories of core issues that warrant research atten-
tion.
First, the legal right to access information focuses on the existing constitutional
and legal conditions in each country. The introduction of new technologies, how-
ever, raises again in many countries the internal political debate about what free-
dom is and what rights citizens should expect. Rights involve several elements: (1)
right to speak; (2) should there also exist a right to be heard? (3) right to informa-
tion access, and the shifting to private information from public information
sources and control; (4) is there a right to receive a response from decision mak-
ers? and (5) what are the ethics ofhaving decision-makers lurk but not respond to
critical Web forums?
264 M.l. Wilson
Second, the legal rights to access are not easily bounded, either legally or so-
cially. Increasingly important is the boundary between public and private in an
electronic world. Restriction of access to information may help or hurt people,
presenting an obligation to balance needs for access and privacy, and to be aware
ofthe information haves and have nots.
Third, rights of access and presentation of information also incorporate lan-
guage/dialect and user community issues. The Internet offers the ability to express
ideas in far more languages and dialects than possible using print or broadcast
media. At the same time, the dominance of English and of a small group of online
languages presents a reduction in choice for access to information.
Harlan Onsrud focuses on one of the important issues over rights in Chapter 18.
He chronicles how the legal context for rights of access to information is changing
in the United States. In particular, Onsrud is concerned about the erosion of access
to public information sources for citizens. The diminishment of legal access comes
from aseries of legislative changes that are often buried deep within legislative
bills. He observes (Chapter 18, 315) that the rights to access public information
gained in the past are slowly being lost as ' ... publishers and government agencies
use the threat of digital technology as an opportunity to limit the rights of citizens
to access information'.
Time
The time dimension cuts across many of the chapters in this book. The ability of
information technologies to reduce or end the friction of distance leaves time as
one of the few remaining baITiers to interaction globally. The importance of meas-
uring and representing accessibility is to permit individuals and institutions to ex-
tract greater value from time, which can be expressed as the currency of the new
economy, albeit a limited and finite resource. Andrew Harvey and Paul Macnab
investigate an important aspect of this theme in Part 11 (Chapter 9), suggesting that
time remains one of the key challenges to interaction now that distance can be
overcome electronically. Using a case study of Canada and its six time zones, the
constraints of interaction are defined clearly by temporal coincidence, and by the
limited windows of real time communication possible at any one time. The socie-
tal increase in types of activities (both virtual and real) subjects individuals and
institutions to allocation decisions.
Processes
The underlying theme of this research direction is to identifY how social, eco-
nomic, and political institutions in different places and times shape access to and
use of information technologies. The importance of institutional players requires
us to understand a range offactors. These include (1) how different types ofinsti-
tutions set agendas for information use and control; (2) how decision making by
Accessibility and Societal Issues 265
these institutions establishes the information infrastructure; and, (3) how relation-
ships emerge between accessibility and social, economic, and political power.
Social processes are also affected by information technologies, challenging schol-
ars to explore the ways that people construct their social networks in an informa-
tion age (a question addressed by Hanson in Chapter 16), and to investigate
whether or not these new technologies require or generate new forms of social
capital. Finally, it is important to seek understanding of how the spatial forces
shaping the use of IT relate to and generate differences and similarities across
places and spaces - a theme acknowledged throughout this book.
15.3 Conclusion
References
Matrix Information and Directory Services Inc. 1999. http://www.mids.org!
MitcheII, W. 1995. City 0/ Bits: Space, Place, and the ln/obahn. Cambridge MA: MIT
Press.
16 Reconceptualizing Accessibility
Susan Hanson
School of Geography, Clark University, 950 Main Street, Worcester MA 01610, USA.
Email: shanson@c1arku.edu
16.1 Introduction
need the resources to do more. Web sites can be packed with facts, but they haven't yet
replaced the old-fashioned, labor-intensive - but more effective - approach offace-to-
face talk. It's hard to champion anational conversation on ignorance. But talking about
what we don't know is an inevitable first step toward breaking down nation-crippling
information barriers.
268 S. Hanson
I begin with this editorial because it highlights what I believe is the focus of this
book - the role of information in access - and it raises (but doesn't really address)
questions about the effectiveness of IT in reducing information inequality and in
increasing access. In drawing attention to local initiatives for increasing personal
interactions between information haves and have nots (all of these initiatives em-
phasize face-to-face contact) and in playing down the role of the Web, this edito-
rial is a cautionary tale for technophiles. I want to use this cautionary tale as a
starting point for rethinking models and measures of accessibility in an informa-
tion age. My goals are
spatially separated. For this reason, accessibility has always been thought of as a
good thing: the more accessibility an individual, group, or area has, the better.
Society worries about those with limited access, and although some observers
have begun to wonder if there might not be such a thing as too much mobility, I
have not yet heard anyone voice the idea that some groups might suffer from too
much accessibility.
Measures of accessibility typically involve counts of the number of opportuni-
ties (number of jobs, square feet of retail space) discounted by distance or some
other measure of impedance. Such measures can be ca1culated for an origin zone,
an individual occupying one or more trip origin points (horne, workplace), or any
number of people whose trip origins have been superimposed on a common point.
In the pre-virtual sense of the word, then, accessibility refers to the theoretical
ability of a person to reach and use dispersed points or zones on aplane (usually
urban space); access measures capture the location of an individual vis avis the
location of a set of potential destinations. Such measures are reaUy measures of
interaction potential. All one needs to realize access, according to this view, is
mobility.
16.3 Silenees
Despite the utility of these measures in revealing access inequalities and inequities
among groups and locations, the measures are limited by their inevitable silences
- the dimensions of accessibility they overlook, neglect, and omit. Traditional
accessibility measures provide rather narrow interpretations of core questions such
as, Do you know what's available at potential destinations? Do you value what's
there? Is it feasible for you to get to these locations and to participate in the activi-
ties there? Do you have social and cultural (not to mention geographic) connec-
tions to facilitate access to those sites?l In traditional accessibility measures, if a
destination is located c10se by, it is considered accessible. As someone at the
Varenius Conference noted, does someone who lives near a library but does not
read or near an airport but does not fly have access? Traditional measures would
answer yes.
Traditional measures of access neglect the fact that people are embedded in
networks of social relations through which information is exchanged, networks
that shape norms and values. That is, traditional approaches do not consider the
informational, social, and cultural dimensions of accessibility. All of these affect a
person's ability and willingness to leave an origin, ability and willingness to trav-
erse distance, and ability to enter and participate at adestination. A job seeker
needs information about current job openings, the locations of those jobs, and the
I Some of these and other silences are addressed in time geography. One example is the
ho urs when stores and other activity sites are open.
270 S. Hanson
Recognizing the central role that information plays in access, some scholars and
policymakers have heralded information technology (especially the Internet and
the Web) as a, if not the, solution to the problem of poor accessibility. Harlan
Cleveland (1985), for example, predicted that widespread availability of informa-
tion would break down baITiers of ignorance and secrecy, eroding hierarchies,
increasing participation, and enlarging democracy. He saw the information revolu-
tion as undermining hierarchies of power based on control, hierarchies of influ-
ence based on secrecy, hierarchies of class based on ownership, of privilege based
on early access to resources, and of politics based on geography. One of the main
ways that IT acquires such power is by erasing the friction of distance, thereby
providing access without mobility. IT is thus seen as the means to overcome the
information segregation that derives in large part from residential segregation.
Stephen Graham (1998) and others (e.g., Hanson 1998) have debunked such
technological determinism as unduly neglecting the lived realities of everyday life
in a material world. In particular, utopian visions of equal access in cyberspace
overlook the pervasiveness and power of the diverse place-based communities in
which the majority of the globe's (and yes, even North America's) citizens live
relatively grounded, even circumscribed, lives, with very real consequences for
access to opportunities. Such utopian visions hint that IT will replace distance-
based (and distance-biased) interactions, diluting their power to shape social life
and ultimately rendering distance and geography obsolete. In my view, such vi-
sions distract attention from the real and difficult job of trying to understand how
IT is complementing and interacting in complex and unforeseen ways with
grounded social, economic, and political exchanges. Precisely because accessibil-
ity requires more than proximity andlor mobility, eliminating the friction of dis-
tance will not yield access.
Reconceptualizing Accessibility 271
Let us assurne that everyone has access to the Web and therefore to almost limit-
less information. What does this mean for accessibility? Shannon and Weaver's
classic information theory (1949) hinges on the idea that information reduces un-
certainty. That is, an item of communication (a word, musical note, phrase, memo,
medical test result) can be considered informative only when it reduces uncer-
tainty. Shannon and Weaver show, moreover, that information's uncertainty-
reducing role is context dependent: an additional musical note in a score reduces
uncertainty only in the larger context of the particular musical passage.
This means that the same piece of information takes on different meanings in
different contexts. Examples abound of words that mean entirely different things
in different contexts. The Catalan word 'prou' means either 'yes, OK' or 'no, not
OK' depending on context and intonation; the essence of Barbara Kingsolver's
new book (1998) turns on a word in Kikongo, 'bangala,' which can mean 'pre-
cious dear' or 'poisonwood tree' depending also on context and intonation. It is
obvious and widely recognized that virtual communication (VC) is not the same as
face-to-face (F2F) communication precisely because it (VC) is decontextualized. 2
The medium of communication and the source of information affect the nature,
quality, and reliability of what's communicated and therefore the degree to which
it reduces uncertainty. Shannon and Weaver's insights about information raise
questions about the information available on the Web. To what extent, and in what
circumstances, is cyber information really informative in that it reduces uncer-
tainty?
I believe that answering this question (and therefore understanding how infor-
mation affects access) will require closely examining how IT intersects with other
forms of communication in place-based communities. In the pre-virtual world,
most people found jobs and most employers found workers through informal per-
sonal contact, not through formal information sources such as employment agen-
cies or newspaper ads. And for good reasons: from the job seeker's perspective,
key information about a potential work site (e.g., what it's really like to work
there) simply does not appear - and never will appear - in formal job advertise-
ments. From the employer's perspective, relying on word ofmouth is cheaper than
using formal advertising outIets and has a higher probability of resulting in the
2 Sproull and Kiesler (1993) have studied how VC changes communication in work organi-
zations and argue that it can lead to more egalitarian patterns of information sharing. But
VC can also lead to more hierarchy and can create more barriers and balkanization (Jones
1995; Van AIstyne & Brynjolfsson 1996). These effects emerge not only because people's
access to IT itselfvaries but also because VC allows us more choice regarding with whom
we interact; it allows each of us to customize our social interactions based on shared in-
terests, thereby narrowing the range of diversity with which we engage. As Van AIstyne
and Brynjolfsson note, however, balkanization is not inevitable; we can 'use IT to select
diverse contacts as easily as specialized contacts' (1996, 1480).
272 S. Hanson
hire of a more productive worker. In each case, the screening function of personal
networks (those of employers and existing employees) increases the likelihood
that a particular employer-worker match is a good one. In short, personal contacts
are effective at reducing uncertainty. Thus studies of labor market processes re-
peatedly find that the information that flows through social networks and everyday
personal interactions plays a pivotal role in shaping people's access to jobs, affect-
ing type of work (occupation, industry), location, and compensation (Granovetter
1974, Hanson and Pratt 1995).
The importance of these personal relations points to the need to recognize how
social and cultural capitaI shape people's access to opportunities. Social capital
'encompasses benefits derived from relations of mutual trust and collaboration; it
thus resides in the relations between members, not in the individuals who compose
it' (Fernandez Kelly 1995,216). Portes and Sensenbrenner (1993, 1323), who are
interested in social capital primarily as it intersects with the labor market, define it
as 'those expectations for action within a collectivity that affect the economic
goals and goal-seeking behavior of its members, even ifthose expectations are not
oriented toward the economic sphere.' They point to four sources of social capital:
(1) value introjection (the socialization into consensually established beliefs), (2)
reciprocity exchanges (the norm of reciprocity in face-to-face interaction), (3)
bounded solidarity (common awareness), and (4) enforceable trust (rewards and
sanctions linked to group membership) (Portes and Sensenbrenner 1993, 1323).3
Cultural capital consists of symbols and values that help people to make sense
of their experiences; it is a byproduct of social capital because it is developed
through the personal interactions that are the raw material of social capital (Fer-
nandez Kelly 1995, 220). Social and cultural capital develop through social ex-
changes that until now have usually and necessarily been largely face to face. The
social and cultural capital available to a person depend on the nature of these per-
sonal interactions, which cannot be ephemeral, fleeting, or singular; social capital
requires repeated contacts and the expectation of on-going interaction. 4
The expectation of repeated interactions points to the importance of geographie
context in shaping social and cultural capital. The webs of social relations that
sustain them develop and grow among the people living in a particular place and
time. Social and cultural capital depend on the sustained contact that comes with
residential rootedness (Hanson and Pratt 1995). Moreover, the geographic content
of the information and knowledge exchanged through these social networks (e.g.,
concerning job opportunities or how to find a good physician) depends in large
3 Fernandez Kelly (1995) and Porte and Sensenbrenner (1995) note that the concept of so-
cial capital has its origins in the dassie sociological texts of Durkheim, Marx, Webber,
and Simmel. Coleman (1988) was instrumental in reviving interest in social capital.
4 A question that bears scrutiny is whether social capital can be built on the Internet. My
suspicion is that the success of IT as a medium for building social capital is likely to de-
pend on one's social position (e.g., gender and dass) and the nature and amount ofsocial
capital one already has.
Reconceptualizing Accessibility 273
part on the geographic extent of the network members' experiences (Hanson and
Pratt 1991, 1995).
This geographic dimension points to another crucial characteristic of social capi-
tal, namely the extent to which it embraces diversity. Do the social networks that
yield social capital connect people across a range of interests and experiences and
across lines of social cleavage (e.g., age, income, class, gender, race)? Putnam
(1993) argues that the civic virtue of social capitallies precisely in its ability to do
so. Putnam's point is that some interests that bring people in a community together
(his examples include bowling in a league, singing in a chorus, or volunteering for
the PT A) will necessarily cut across other lines of social cleavage (income, gen-
der, religion, class, etc.). Fernandez Kelly's study of teen motherhood in a Balti-
more ghetto suggests how sei dom such cross-cutting ties are in fact part of the
social capital of the poor.
Note the importance of embodiment in this argument: bowlers or singers may
share a passion, but it is precisely their coming together in the flesh, as whole per-
sons, that brings people in contact with difference and therefore bridges lines of
social cleavage. Joining together with like-minded souls on the Internet to pursue
a common interest does not necessarily have the same result. What is the probabil-
ity that a person selected at random from one's social network will be the source
of new or surprising information about employment opportunities? Are the people
who supply one with information about jobs all working at the same level in the
same kind of jobs? If so, any job information received from them is likely to be
less useful for social mobility than if that information came from a network of
diverse sources (Granovetter 1982, Hanson and Pratt 1991). Do networks extend
beyond the immediate community, or are they socially and spatially confined? The
amount of diversity that is built into the social capital of a place or person or group
crucially affects accessibility. Because geographic mobility allows people to sepa-
rate themselves spatially from those who are different, one serious source of un-
equal access in metro areas is the relative homogeneity - and impoverished
homogeneity - in the social and cultural capital of the poor.
In sum, access to jobs requires more than having proximity/mobility and even
more than possessing the needed human capital (education, skills, experience) for
a particular type of employment. It requires information, but more importantly, it
requires certain kinds of information - the kinds that inhere in social and cultural
capital. Traditionally social and cultural capital have developed through networks
of F2F contact, in which information has been contextualized and a basis for trust
established. In thinking about accessibility in an information age, how might IT
intersect with these often place-based and place-biased information networks?
How might IT be used to intervene strategically to increase the access of those
who currently lack it?
274 S. Hanson
I have argued that traditional measures of accessibility are silent on the role of
information and that any information-age concept or measure of accessibility must
incorporate information, virtual and grounded, electronic and F2F. In this final
section, I first sketch out how the pre-virtual information (exchanged F2F) em-
bedded in social and cultural capital already does and might in the future interact
synergistically with virtual or cyber information. Second, I consider desiderata in
an information-age accessibility measure.
Some have seen IT as the perfect answer for those whose social capitallacks the
diversity needed to connect them with 'good' opportunities, such as good jobs.
Because most people find out about jobs through personal contacts, people with
highly localized social networks are unlikely to hear about jobs that are located
outside the immediate community.5 Moreover, because previous co-workers are
an important source of information about new jobs (more so for men than for
women), having held few or no jobs in the past also constricts one's information
about job opportunities. IT has been proposed as an ideal way to obviate these
problems and supply people with the job information they need.
But this suggestion shows no appreciation for why social networks are so popu-
lar as conduits of job information in the first place and why digital job banks have
not been very popular with either workers or employers. People value F2F infor-
mation from known sources: the on-going nature of a social relation enables trust
and sheds light on the veracity of the information; in this sense the source acts as a
screen. The importance people accord the information they exchange face to face
underlines the crucial role of context; information exchanged electronically means
something different from information exchanged in person. Yet despite the power
of F2F, the Web does have enormous potential for disseminating job and em-
ployment-related information and especially for bridging the divide between in-
formation haves and have nots. This suggests the possibility of a productive union
between grounded and virtual information exchanges.
I've recently begun a study of entrepreneurship that focuses largely on how
business start-ups and self-employment are related to and embedded in people's
labor market experiences. A couple of anecdotes from entrepreneurs we've inter-
viewed in Worcester illustrate how these people are combining IT with F2F to
increase access. The point here is not so much that lots of small business owners
are using the Internet; relatively few now do, although many more voice plans to
5 Certainly not all social networks that connect people with jobs are localized around par-
ticular workplaces. Meir and Giloth (1985) found that word-of-mouth recruiting explained
high unemployment and long commutes to low-wage jobs among Mexican Americans in
a Chicago neighborhood that had good local job opportunities. The employers in this
neighborhood used word-of-mouth recruiting, but their existing employees did not live
locally and had no local networks. Neighborhood residents did not, therefore, have the
needed network access to jobs in their own neighborhood.
Reconceptualizing Accessibility 275
use the Internet soon. What is striking is how use ofthe Web combines in interest-
ing ways with personal contacts to serve the business owners' interests.
One woman, whose home-based business is marketing the products of high-tech
firms to scientists and engineers, uses the Web extensively. A considerable share
of her c\ients are in California and Texas. When we asked how she had penetrated
those markets, she cited word of mouth and her firms' Website, two methods that
she c\early sees as complementary. Having worked in the high-tech field for seven
years before launching her own company, this woman had an extensive array of
personal contacts among potential c\ients before start-up She now prepares a quar-
terly newsletter on marketing, which she faxes to all of her current c\ients (most of
whom came to her via word ofmouth) and to potential c\ients whom she's gleaned
from the Web as weIl as from word of mouth with existing c\ients and others. The
news letter refers readers to her Website, which then prompts some personal con-
tact and yields new c\ients.
A second example - of a woman who runs a flooring company - suggests how
business owners use the Web strategically to broaden and diversify information
sources weIl beyond what would be available through place-based personal con-
tacts. When asked whom she relies on now for information and advice in running
her business, she replied succinctly (and quite distinctly): 'My computer.' She
uses the Internet extensively to learn what her competitors are doing as weIl as to
identify potential c\ients. She has, in fact, been able to extend her market area by
learning from the Web about potential jobs, hut her floor installers have been will-
ing to travel to these distant locations only because (and when) they have personal
contacts (family, friends) there. These examples illustrate how small business
owners weave together various forms ofIT with F2F to promote their businesses.
Here are a few additional examples, with a more futuristic flavor. One involves
setting up computer work-stations and Internet links in study rooms/labs for chi 1-
dren in public housing projects or in neighborhood study centers in low-income
areas. My sense is that most people rely upon their social networks in learning
new IT and in negotiating the Internet; people often, for example, visit new web
sites because they have heard about them in F2F exchanges. These facilities will
need to be staffed with people who can help the children use the technology, a
F2F /IT interface in itself, and offer them personal contacts that link them to the
world beyond their immediate environments. Such adults might be college stu-
dents, teachers, or retired persons whose experiences straddle the local community
and other places and who can personally connect the children to opportunities out-
side their immediate experience. The synergistic nature of IT and F2F interaction
will be evident, for example, in that staff in these facilities will often use the Web
to identify opportunities previously unknown to themselves.
In the context of job search, it seems possible that intermediaries such as teach-
ers, members of the c\ergy, volunteers, or members of non-profits could play an
important bridging role. Perhaps both job seeker and intermediary would have
access to job information on the Web and the intermediary could act as go-
between with employers, helping to screen applicants. In this way IT might also
he\p match teenagers with employers in internships and apprenticeships that
276 S. Hanson
would link them into work-based social networks. IT will be most powerful in
expanding access if it enriches social capital by broadening and diversitying in-
formation exchanged face to face and by prompting the discussion of new ques-
tions.
How might information-age accessibility measures incorporate these informa-
tional dimensions of access; i.e., how might accessibility measures include aspects
of access other than simple opportunities-discounted-by-distance? I suggest three
components as desiderata for information-age measures of access. First, I think
that we should not neglect the spatial arrangements captured in good old-
fashioned spatial accessibility measures; these will remain important and should
be retained in information-age accessibility measures. Second, as many other au-
thors in this book have suggested, measuring access in the information age means
including one or more measures ofthe access ofpeople and places to IT itself - in
the horne, school, neighborhood, or workplace - and to locations in cyberspace.
Third, access measures must incorporate measures of the sociallcultural capital
that individuals, groups, and areas have access to. This means devising measures
that capture the collective information assets - and especially the diversity present
in these group assets - that network members can tap into.
Operationalizing these concepts presents achallenge, but with GIS, devising
such measures is not out of the question as long as the U.S. Census long form re-
mains. For example, one could use the journey-to-work files derived from the long
form to characterize for each person or household or area Cl) the occupations and
industries (measured by three-digit census codes) of employment and the wages
earned per worker in that person's or household's place ofresidence and place of
work, where places could be census tracts, block groups, or blocks and (2) the
variance in occupations, industries, and wages by place of residence and place of
work. The residence-focused measures could be weighted by length of residence.
These are rough measures, to be sure, but the data are readily available and they
do capture key aspects of social and cultural capital, i.e., the embeddedness of
people in particular milieus. Another possibility that does not rely on the census
long form but instead uses data from employers lies in examining the zip codes of
residence of the current employees in a given work place (where place can be de-
fined as an employer/establishment or as the collection of employers in an area).
Mapping these would shed light on the spatial extent of information exchanges
that people have access to at the workplace. I offer these as surrogate measures for
the nature and diversity of information exchanged F2F through social networks.
Reconceptualizing Accessibility 277
16.7 Conclusion
References
Cleveland, H. 1985. The twilight ofhierarchy: Speculations on the global information soci-
ety. Public Administration Review 45(1): 185-95.
Coleman, J. 1988. SociaJ capital in the creation of human capitaJ American Journa/ 0/Soci-
ology Supplement: S95-S120.
Fernandez Kelly, P. 1995. Social and cultural capitaJ in the urban ghetto: Implications for
the economic sociology of immigration. In Portes, A. (ed.) The Economic Soci%gy 0/
Immigration: Essays on Networks, Ethnicify, and Entrepreneurship. New York: Russell
Sage Foundation, 213-47.
Graham, S. 1998. The end of geography or the explosion of place? Conceptualizing space,
place, and information technology. Progress in Human Geography 22(2):165-85.
Granovetter, M. 1974. Getting a Job: A Study o/Contacts and Careers. Cambridge: Har-
yard University Press.
Granovetter, M. 1982. The strength of weak ties: A network theory revisited. In Marsden,
P. and Lin, N. (eds.) Social Structure and NetworkAna/ysis. Beverly Hills: Sage. 105-30.
Hanson, S.. 1998. Off the road? Reflections on transportation geography in the infromation
age. Journal o/Transport Geography 6(4):241-9.
Hanson, S. and Pratt, G. 1991. Job search and the occupational segregation ofwomen. An-
na/s o/theAssociation 0/American Geographers 81:229-53.
Hanson, S. and Pratt, G. 1995. Gender, Work; and Space. New York: Routledge.
6 We might consider also how people use IT not only to increase their access to information
and opportunities but also to control and limit the access of others to the self (e.g., via
turning off the telephone, using caller ID and answering machines, checking e-mail spo-
radically).
278 S. Hanson
Jones, S. 1995. Understanding community in the information age. In Jones, S.G. (ed) Cy-
bersociety: Computer-mediated Communication and Community. Sage publications.
Kingsolver, B. 1998. The Poisonwood Bible. New York: Haprper Flamingo.
'Information Inequality,' Boston Globe August 23,1997.
Meir, R. and Giloth, R. 1985. Hispanic employment opportunities: A case of internaIlabor
markets and weak-tied social networks. Social Science Quarterly 66:296-307.
Portes, A. and Sensenbrenner, J. 1993. Embeddedness and immigration: Notes on the social
determinants of economic action. American Journal ojSociology 98: 1320-50.
Putnam, R. 1993. The prosperous community: Social capital and public Iife. The American
Prospect No. 13 :35-42.
Shannon, G. and Weaver, W. 1949. The Mathematical Theory ojCommunication. Urbana,
IL: University of IIIinois Press.
Sproull, L. and Kiesler, S. 1993. Computers, networks, and work. In Harasim, L. (ed)
Global Networks: Computers and International Communication. Cambridge, MA: MIT
Press.
Van Alstyne, M. and Brynjolfsson, E. 1996. Could the Internet balkanize science? Science
274:1479-80.
17 Revisiting the Concept of Accessibility:
Some Comments and Research Questions
Sylvie Occelli
IRES - Istituto di Ricerche Economico Sociali dei Piemonte, Via Nizza 18,
10125 Turin, Italy. Email: occelli@ires.piemonte.it
17.1 Introduction
Almost all urban systems in developed countries are undergoing a number of insti-
tutional, socio-economic and cultural changes, pushing them towards a 'new' so-
cietal configuration that is generally taken to be more democratic, better educated,
culture-based, and environmentally sensitive, i.e., the so-called Post-Fordist soci-
ety (AminI994). Space-adjusting technologies, and particularly the New Informa-
tion Technologies (NIT), are playing a substantial role in this transition, since they
affect both the range and time-related organization of activities offered in an urban
setting, as weil as the ways in which individuals participate in them (see Castells
1989, Graham and Marvin 1996).
Due to its intrinsic ability to provide the 'connections' between the pattern of
activities and their interdependencies, accessibility is a very sensitive concept for
analyzing these changes. In the current transition to a Post-Fordist society, how-
ever, the notions traditionally used reveal many drawbacks. Arevision is needed
to (a) improve understanding of the space-time changes taking place in modern
cities (Couclelis 1996, Bertuglia and Occelli 1997) and (b) provide appropriate
indications for coping with practical planning problems and, in particular, policies
relating to accessibility.
While not exhaustive, this paper identifies a number of aspects likely to consti-
tute the accessibility question in Post-Fordist urban development (see Rabino and
Occelli 1997). The discussion builds upon recent work in which the need for this
revision has already been advocated (Occelli 1998a). It was argued that, at least on
substantive grounds, a major area for exploration is the impact of New Informa-
tion Technologies (NIT) on the time dimension underlying the notion of accessi-
bility. It was pointed out that the NIT gives accessibility a new role and new po-
tential, making it not simply a time-space opportunity, but also aresource. This
implies going beyond conventional definitions of accessibility and requires a
broader perspective of analysis.
One implicit aim for this chapter is to show how reasoning about accessibility
(i.e., developing an analytical framework, such as a modeling activity) can help
both in disentangling accessibility problems and in defining better policy meas-
280 S. Occelli
ures. The discussion is divided into four parts, which form the main building
blocks ofthis reasoning.
The first part revisits the concept of accessibility, emphasizing how it relates to
the main components of a spatial system, and the second part recalls classical
definitions of accessibility proposed in the literature. It is argued that these reflect
the features of the city that have emerged during the historical process of urbani-
zation. A meta-typology of urban development is proposed as a way to improve
understanding of accessibility in relation to the stages of urban evolution.
The third part shows that extensions in the analysis of accessibility do not result
only from phenomenological issues associated with the transition to a Post-Fordist
society. They are also related to a shift in analysis within the mainstream of quan-
titative geography. Implications of this new approach to the understanding of ac-
cessibility are illustrated.
The last part focuses on the formulation of some of the issues likely to be given
priority in future research on relationships between NIT, urban evolution and ac-
cessibility. In this connection, a research project on accessibility in the Turin Met-
ropolitan Area ofItaly is described.
1 From a different point of view, the junction role of accessibility has already been recog-
nized in New Urban Economics, where accessibility has been interpreted as an extemal-
ity, thus sharing the properties of a public good (see Papageorgiou 1987).
Revisiting the Concept of Accessibility 281
Accessibility can thus be considered as the outcome and, at the same time, a
component of the interaction process underlying the functioning of urban systems.
It is intuitively evident, ceteribus paris, that the greater the importance given to an
urban product and the lower the constraints on its enjoyment, the higher its acces-
sibility. On the other hand, a high level of accessibility is likely to enhance the
importance of an urban product and stimulate the demand, thus increasing its po-
tential for interaction (i.e., the level of flows to and from the area where the prod-
uct is located, and the number of contacts between people).
Empirically, the 'junction role of accessibility has long been recognized in land-
use and transportation analysis. Several measures of accessibility have been for-
mulated and provide meaningful links within and between residence- and organi-
zation-based urban sectors. These relate the various activities available to indi-
vi duals and organizations with the systems of transport and communications,
which allows individuals to overcome the impeding distances and to participate in
specific activities (see Hansen 1959, Wilson 1971, Morris Dumble and Wigan
1979, Leonardi 1979, Wachs and Koenig 1979, Koenig 1980, Hanson 1984, Han-
son and Schwab 1987).
More recently, this intrinsic junction feature of the concept has been exploited,
and measures of accessibility are explicitly included in the performance indicators
developed for the analysis and evaluation of spatial structures (Clarke and Wilson
1987a, 1987b, Bertuglia, Clarke, and Wilson 1994). The junction role of the con-
cept is, however, also responsible for certain elusiveness underlying the notion of
accessibility. It explains the ambiguity, which often accompanies the common use
of the term, associated primarily with travel-derived demand. Thus, accessibility is
used to mean both the proximity and ease of interaction and the possibility of in-
teraction.
As far as the former is concerned, its application in the field of transportation
and land-use planning has generally involved an interpretation connected in some
way with the spatial separation of human activities, usually expressed as a physi-
cal distance or a time. Time as a locational and co-locational continuum thus rep-
resents a basic dimension of accessibility (Car1stein, Parkes, and Thrift 1978).
In relation to the latter, the role of time associated with spatial separation makes
it possible to relate the notion of accessibility to the field of choices of interaction
available to an individual (Weibull 1980). The extension of this field depends on
the individual's capacity and resources, and also on the patterns of opportunity he
or she possesses in relation to the mix of activities, and their functional and spatial
282 S. Occelli
While making the junction role of accessibility more complex, this endows it
with new potentialities. Accessibility is no longer just a property to take advantage
Revisiting the Concept of Accessibility 283
of, but a resource that a city as, a collective social entity, makes available 2• It is
the increased potential in its junction role that makes accessibility so relevant in
the new information society. Although not yet fully recognized, this is foreshad-
owed in a number of re cent studies where accessibility is referred to as a meaning-
ful analytical notion for investigating the impact of innovation on the functional
and spatial organization of urban activities. For example, Kobayashi, Sunao and
Yoshikawa (1993), use an accessibility representation to describe the interdepend-
encies created by knowledge exchanges. Martellato (1993) considers transport and
telematic accessibility as distinct inputs to the production function of firms. He
analyzes how changes in the price of such services impact on the locations of
firms. Bertuglia, Lombardo and Occelli (1995) develop a model for simulating
urban scenarios in which the location choices of firms depend on both transport
and telematic accessibility.
On methodological grounds, acknowledgement of the hard and soft components
of NIT changes our way of conceiving accessibility and hence its role in relation
to Post-Fordist urban development. As discussed later in this chapter, there can be
no single notion of accessibility. There will be many different definitions and
measurements of accessibility (see IRES 1995, Kwan 1998 and Helling 1998).
Depending on how it is represented, the differing knowledge of accessibility held
by different individuals mirrors the complexity ofthe concepe.
The implications associated with the junction role of accessibility have never been
fully explored in the literature. Nonetheless, they have inspired a number of defi-
nitions, which have shed valuable light on many urban phenomena and spatial
2 It has been po si ted (Bertuglia and Occelli 1997) that this soft component is by far the
most important aspect of present-day innovation. It can be related to economically based
notions of knowledge, experience and learning, or to the circuits of communicative inter-
action that form the social fabric; it can even be associated with the so-called software
network (including education, the arts, and science) that embellish human infrastructure
(Andersson et al. 1993). More importantly, however, by changing interaction patterns, it
can affect a whole range of behavioral (i.e., goal seeking, explorative, imitative, hyper-
selective) processes. These can act at both the individual and collective level and affect
spatial systems in a number of different ways to feed its evolution. One relevant implica-
ti on is that everyone can be an agent of innovation. By pointing at the ordinary agent -
whether an individual, household, organization, or collective society - as repository of
innovation, we have a less elitist view.
3 Acknowledging the individual ordinary agent (person, household or organization) as the
main repository of innovation, implies that hislher notions of accessibility are increas-
ingly important - how they are derived from individual agents' representations of their
daily activity patterns and how these representations are continuously adjusted, modified,
and re-created.
284 S.Occelli
Table 17.1 lists a few of these definitions and, although not exhaustive, gives an
idea of the evolution of the accessibility concept. Also mentioned are the main
approaches to urban interactions from which the definitions are derived or to
which they can be related.
Revisiting the Concept of Accessibility 285
The proposed typology is, of course, highly simplistic and ignores other deter-
minants of urban change or, more importantly, continuities over time. It is how-
ever detailed enough to show that, in the evolution ofurban systems, an increasing
complexification of cities has taken place. This is particularly evident when we
contrast the descriptions ofFordist and Post-Fordist types ofurban development.
The Fordist City is generally seen as a place of industrial production, whose
evolution is driven by the growth of (export oriented) industrial sectors, which in
turn trigger the growth of the resident population and services. These services in-
clude all urban activities that are oriented to the local market (i.e., the resident
population of the city). The spatial interactions consist mainly of flows of employ-
ees who travel between the many places of residence to the few places of work.
The structure of the city is intrinsically stable and its path of growth is continuous
and unvarying. The most relevant determinants of accessibility are associated with
the costs of commuter movements and the opportunities within zones (i.e., the
range ofurban activities and vacant land), and with transportation (e.g., mass tran-
sit, road capacity, and speed). By contrast, the Post-Fordist city is seen as an in-
formation-based spatial system in which the kind and characteristics of relation-
ships become increasingly important (Rabino and Occelli 1997).
Not only are urban relationships more numerous and varied (by type, time inten-
sity, frequency, and number of actors involved), but they are also evolutive and
self-organizing, based on different spatial and temporal scales (the world-wide and
local) and interacting with the relationships of other sub-systems (e.g., the envi-
ronmental and socio-cultural sub-systems, etc.). The most salient feature of the
Post-Fordist City is, therefore, the pattern of networking relationships. Not unex-
pectedly, accessibility has a multi-dimensional nature related to the interaction
opportunity (which can be social, functional, physical, and virtual) of individuals
and organizations, resulting from their capacity to enter various fields of urban
interaction.
From the definitions of accessibility in Table 17.1, none of them are associated
with the pre-industrial city. While the definitions rooted in the physical determi-
nistic and economic-functionalist approaches can be related to characteristics of a
Fordist city, those derived from the economic-behavioral, spatio-temporal, and
informational approaches clearly allude to a Post-Fordist type of urban develop-
ment. The so-called informational approach, in particular, accommodates the role
of the New Information Technologies in our understanding of the present-future
continuum and ofplanning for the future (Miles and Robins 1992).
Tables 17.1 and 17.2 recognize the increasing complexification of cities. That
accessibility, too, is becoming more complex is also acknowledged. As noted, the
time-space shrinking possibilities and enabling potentials associated with NIT, are
major determinants ofthis complexification.
Revisiting the Concept of Accessibility 287
Pre-industrial city
FEATURES (merchant and agricul- Fordist city Post-Fordist city
tural based society)
Socio-economic and institutional aspects
Production sectors Agriculture, electrical Cars, armaments, Computers, capital
machinery, steel ships consumer durables, goods, optical fibers
petro-chemicals. Mass telecommunications,
production
Tertiary sectors Domestic services, state Growth of social and Expansion of infor-
and local bureaucracies, financial services. mation services. New
growth oftransportation Decline of domestic forms of craft
and distribution services production
In/rastructures Canals, railways, roads Electrical cables, Digital communica-
highways, airlines, tions network,
airports satellites
Social organiza- Rigid c\ass divisions. Unified c\ass Pluralistic c\ass for-
tion and popula- Urbanization, high formation and parties. mations, multi-party
tion trends population turnover Concentration of system, regional di-
population in urban versification. Counter
areas urbanization and
aging of population
Aspects 0/regimes Craft unions and early Welfare state and its New-style participa-
0/ regulation sociallegislation crises tory decentralized
welfare state
Spatial aspects
Settlement pattern Isolated, small settle- Formation of polar- Metropolitanization,
and urbanization ments with stable popu- ized, high-density edge cities, dispersed,
I,processes lation agglomerations. polycentric settle-
Marginalization of ments ofvarious size,
peripheral areas network of cities
Type 0/ interaction One-to-one. Open non One-to-many. Radial Many-to-many. Inter-
connected network network connected network
Determinants 0/ Physical distance and Cost of movement, Interaction opportuni-
accessibility transport. Place-based centrality, transport ties, physical vs. vir-
determinant means. Person-based tual interaction. Field-
determinants based determinants
Major urban Housing and health Employrnent, cost of Environmental sus-
issues conditions opportunities, re- tainability, quality of
source allocation , life, urban perform-
urban growth ances, city competi-
ti on
288 S. Occelli
The discussion so far has provided arguments on the need for revisiting notions of
accessibility. However, this need does not result solely from phenomenological
issues that mirror the transition to a Post-Fordist society. Increased complexity in
the notion of accessibility also depends on a broader shift in the approach to
analysis. This is related to several changes in the modeling field (and more gener-
ally in quantitative geography) over the past twenty years. Some ofthese changes,
which are also major topics of enquiry in the Varenius Project, have been ad-
dressed more extensively elsewhere (see Rabino and Occelli 1997) and are sum-
marized below.
(1) A first source of change sterns from the epistemological background. The
acknowledgement of limits to rationality and the need to develop a new phi-
losophy for social action has fostered an interest in the cognitive interpreta-
tion of modeling. In this approach, the model is a means for hypothesis test-
ing, targeting ill-defined problems and yielding alternative visions of likely
futures. This is distinct from the structuralist interpretation of models that
seek a more rigorous understanding of the workings of the system. Whereas
differences between structuralist and cognitivist interpretations are becoming
more noticeable, the complementarily of their roles in dealing with urban
phenomena is also more evident. If, for complex systems, such as cities, 'the
multiplicity of disciplinary viewpoints and paradigms is the norm' (Batty and
Xie 1996, 202), then one major challenge for urban modeling and geographi-
cal analysis is how to reconcile the connections between the two interpreta-
tions. In this direction, some possibilities lie in:
(2) A second source of change is the new technological backcloth resulting from
the introduction of NIT and, in particular, from the increasing power of desk-
top computing. Three related consequences are:
Revisiting the Concept of Accessibility 289
(3) A final source of change concerns the socio-cultural context. As the cultural
and information levels of society as a whole are rising, the socio-cultural con-
text is becoming more demanding and selective in the kind of knowledge ex-
pected (Knight 1995). First, a new awareness is emerging about the multiplic-
ity of processes that combine to produce overall changes in urban systems.
The conceptual unity of the urban system is no longer considered an axio-
matic entity. Rather, it results from a continuous re-definition that gives em-
phasis to the interplay of knowledge-driven actions (behavior) of a variety of
actors. Planning questions also need to be put in a new perspective. For in-
stance, besides the need to disentangle the key questions to be answered by
policies, new needs are emerging for devising pro-active policies to anticipate
problems A
4 A different way to conceive relationships between the observer (the analyst) and reality
(the modeled system) is advocated. The urban model er is part ofthe observed reality and,
as such, he or she is an agent of urban change like any other agent. Furthermore, his role
as a maven no longer holds. On the one hand, as a problem-solver his role does not differ
from that of any other practitioner. On the other, as a problem-definer, the role is likely to
be enhanced. Modeling activity (in both the cognitivist and structuralist domains) makes it
possible to set up intelligent interfaces based on NIT that favor the comrnunication ofthe
various system descriptions and to arrive at a collective shared description. This idea fol-
lows arguments in policy analysis and management science that point to the need to link
290 S. Oeeelli
The circular links between these elements (indicated with Arabic numbers in
Figure 17.1) relate to the process of abstraction underlying any approach to acces-
sibility. The cross-links, labeled 'validation and experimentation' refer explicitly
to specific features of the modeling activity. Although the methodological under-
pinnings of the contents of Figure 17.1 require more detailed discussion, this pres-
entation addresses only the main differences between the two approaches and their
likely implications in extending the accessibility concept.
action and organizations more effeetively to enhanee the capacity of loeal planning insti-
tutions (see Bennett and MeCoshan 1993, Bryson and Crosby 1998).
Revisiting the Concept of Accessibility 291
sibility models. Third, the 'observable' is also different in the two ap-
proaches. In the conventional approach, accessibility is assumed to be physi-
cally determined by the spatial and functional properties of the system. How-
ever, in the unconventional approach, accessibility is the outcome of a
multiplicity of changing individual perceptions - collectively shared and
evolving entities that may appear as policy issues. Finally, as far as the men-
tal models are concerned, 'awareness and consciousness' are features in the
unconventional approach, which seem unaccounted for in the conventional
approach.
• A second set of differences lies in the links. It is assumed that encoding of the
understanding process begins with an 'observation activity'. As shown in
Figure 17.1, the analytical path moves clockwise in the conventional ap-
proach (observables -7 theories -7 mental models -7 system models) and
anti-clockwise in the unconventional approach (observables -7 system mod-
els -7 mental models -7 theories). In then unconventional approach greater
emphasis is given to the shared descriptions of models (the so-called negoti-
ated accessibility) that result from an analytical activity (experimentation) to
connect the representations that underlie 'negotiated accessibility' and indi-
vidual mental models.
• A last point relates to transformation in the activities of experimentation and
validation in the unconventional approach. Experimentation becomes an ac-
tivity aimed primarily at seeking agreement between the representations of
accessibility held by individuals in achanging environment and the 'notions
of accessibility' that have been agreed upon. Validation does not necessarily
need 'physical accessibility' to be carried out. It relies on a kind of pseudo-
scientific process based on the relationships established between 'negotiated
accessibility', theories and system models, in which societal issues (e.g., eq-
uity, sustainability, and quality oflife) set the fundamental yardstick.
Figure 17.1 posits that, even from a strict1y methodological point of view, acces-
sibility is a manifold concept. Three distinct analytical dimensions for represent-
ing the notion of accessibility are suggested:
........ ....
nbstrAccions
under lying our
un dtr Ilwd in g
~~_I_O_",_._I<_h~n_'OO__.b__~
mp/ememaflOIt V.lodauon
The ob.en'abl .. : lho
ntgotiJlted
8CC' ibilily
Socwl PI'OCeJoSeI
COllcepmal proces..'oe.'"
~
lan ofthe encodmg process
Figure 17.1. Approaches to the analysis ofaccessibility (adapted from Turoff 1997)
From a methodological standpoint, there is a role for NIT (particularly for its
'soft' component) in creating, updating, and innovating these threefold representa-
tions of accessibility. As suggested elsewhere (see Bertuglia and Occelli 1997),
NIT's effectiveness is strictly associated with the awareness (perception, informa-
tion, and knowledge) of innovation potential (i.e., the technological self-
referentiality mentioned by Sui (Chapter 7). This me ans that the expected impacts
of NIT on these representations are likely to have no lesser role than the appear-
ance ofNIT itself(i.e., the hard component ofNIT).
A major question concerns the potential of NIT in making it possible, and eas-
ier, to link the representations with different analytical dimensions. This extricates
us from a discussion of complexity criteria (i.e., by improving our evaluation of
the unpredictability and discontinuity of urban changes) and moves our attention
to aspects likely to have many practical implications, such as those regarding the
'reducibility question'. To deal with this question, we make two assumptions:
uisite variety', often mentioned in the literature (see Friend and Jessop 1969,
Casti 1986, Batty 1995).
In this paper the importance ofthe 'junction role' ofaccessibility has been empha-
sized, bridging the spatio-temporal and spatio-functional component of spatial
systems. To cope with the space-time changes occurring in the transition to a Post-
Fordist type of development, it is suggested that this junction role should be revis-
ited and our notion of accessibility extended. An effort has been made to justify
this extension on conceptual, phenomenological, and methodological grounds.
The role of NIT has also been addressed. From all these points of view, the in-
creasing complexity of the concept of accessibility has emerged as a main justifi-
cation for its extension.
This does not mean that the conventional notions of accessibility no longer ap-
ply. It simply reflects the fact that accessibility is an intrinsically manifold notion,
encompassing definitions that co-exist, but are not reducible to each other. In par-
ticular, three distinct analytical levels were mentioned as having relevance for
representing accessibility - the individual, systemic, and policy levels. Although
this may appear a trivial result on speculative grounds, it is certainly not simple
from a policy point of view. Most definitions of accessibility currently used do not
consider this distinction and usually assurne that the same indicator of accessibil-
ity can be applied and have the same meaning in very different planning contexts.
In dealing with methodological aspects, it was emphasized that the many differ-
ent notions of accessibility also depend on the kind of representation we have. It
was argued that NIT could have a substantial role in the formation and updating of
representations and it was suggested that NIT, in particular its soft component,
might be helpful in allowing 'informational convergence' between the different
representations of accessibility, thus providing a bridge between the various ana-
lyticallevels.
Recent research suggests a number of questions and issues that need considera-
tion (see Couclelis 1996, Handy and Niemeier 1997, Helling 1997). Building upon
suggestions in Occelli (1998a), the following issues and questions are identified
for future research:
(1) The first relates to the difficulty in identifying the appropriate level of defini-
tion for a given representation of accessibility. The problems are both concep-
tual and empirical. Conceptually, the need to formulate an appropriate time-
space frame of reference has been pointed out (see Holly 1978). Also, the
kind of description of the 'effort necessary to get at an urban product' at the
individual and systemic levels can be quite different as a consequence of the
ways distance and travel impedance are incorporated conceptually and em-
Revisiting the Concept of Accessibility 295
pirically (see Couclelis 1996, Handy and Niemeier 1997). Morris, Dumble
and Wigan (1979) identified empirical problems in constructing of indicators
whenever the two temporal dimensions of accessibility (the locational and co-
locational continuum and the 'ladder' of time are taken into account. Do we
wish to examine the opportunities available (through process indicators) of
what is potentially offered for different types of individuals, independently of
their observed behavior? Or, do we wish to focus on the properties of out-
come indicators, such as observed journeys to activities, travel times for jour-
neys-to-work or for leisure, and so forth? What should be investigated by
means of process indicators at one spatio-temporal scale, however, would re-
quire outcome indicators at another. Different spatio-temporal scales reveal
different issues, which can have contrasting implications for policy. A crucial
point in this respect is the management or control of the so-called 'border ef-
fects' - variations in accessibility betweendifferent time-space levels. For ex-
ample, improving regional accessibility in an international context will not
necessarily have positive effects on all local areas. Conversely, higher acces-
sibility levels in a local area do not necessarily guarantee the improvement of
its connections with regional or international markets.
(2) A second set of questions is raised by two major features of 'Post-Fordist' type
urban development, namely the growing importance of interactions (the de-
mand for mobility and communication) and the effects that such interactions
may have on sustainability and the quality of life. An important implication is
that any definition of accessibility should be accompanied by an evaluation of
the associated benefits accruing to the individual, an organization, or the city
as a whole. In particular, considering accessibility as a resource implies that
attention should be paid to:
• the type and quality of urban products, such as the various kinds of activ-
ity relative to their temporal organization and location. This means that
'what' is to be accessed and also 'how' are relevant in determining the
benefits. This in turn raises questions relating to scarcity and efficiency
in the provision of a range of urban services, as weil as the co-ordination
of the different ac ti vi ti es (e.g., the opening times of services and trans-
portation availability);
• the ways of overcoming spatial separation, such as the kinds of commu-
nication links involved in various human interactions. In this connection,
attention needs to be given to the 'value' of time associated with the
communication links, integrating the analysis of both cognitive and prac-
tical aspects ofvaluations;
• the kind of trade-offs likely to be involved: the positive effects at the in-
dividual level and any negative externalities at more aggregate levels
(e.g., increased traffic congestion in some areas of the city). One further
implication, particularly relevant in planning, concerns the relationships
between accessibility, mobility, urban form (the spatial distribution of ac-
tivities and patterns of land use), and the environment. A less myopic
296 S.OcceJli
(3) A further set of issues concern the representation of accessibility and the
recognition that the kind of 'knowledge' held by individuals and decision-
makers about accessibility is a fundamental determinant in the use of the ac-
cessibility resource, as weil as in its preservation and regeneration. In the cur-
rent transition to a Post-Fordist urban development, there is a risk that gaps in
the different representations of accessibility (particularly between those of the
general public and decision-makers) might exacerbate accessibility needs,
raising problems of equity and social justice. Improving information about
accessibility, therefore, should be an essential component of any policy strat-
egy, since it can improve not only accessibility, but also social equity.
This last concern is a major focus of the lRES survey. The purpose of the
project is to answer two main questions which, due to the provisions set out in
anational law on local govemment enacted in 1992, are also relevant from a
policy point of view: (1) how is accessibility perceived by the residents in
metropolitan area? (For instance, what knowledge do people have oftheir ac-
cessibility and how do they value it?); and (2) what is likely to be the social
acceptability of alternative bundles of accessibility measures? The underlying
thesis of the project is that knowledge about accessibility, and the representa-
tions that individuals and decision-makers have of accessibility, is likely to be
no less significant than the accessibility policy measures themselves.
Building upon observations made previously, Figure 17.2 shows a concep-
tual framework for defining the questionnaire 5 • Two main levels of definition
of accessibility are considered: (l) a systemic level, and (2) an individual
level. Accessibility at the systemic level reflects the functional and spatial or-
ganization of the activities provided by various transport agencies and service
authorities (the mix of opportunities, trave1 times, transport services, opening
5 The IRES survey is a pilot survey for a larger research project on accessibility in the Turin
Metropolitan Area. The survey was conducted in November and December 1998 by
means of horne interviews. The individual is used as the unit of analysis. About 400 per-
sons were interviewed. Due to resource constraints, the study area was limited to the
western sector of the MA. The questionnaire was divided into three sections. Section I
covers basic information about the individual and hislher family. Section 2 covers infor-
mation about the action space of the individual in hislher urban environment. The respon-
dent was also asked for an assessment of this action space. Section 3 requested an evalua-
tion of a set of alternative measures that could be introduced to improve accessibility.
Revisiting the Concept of Accessibility 297
Long-run
Representations
0/ accessihilifJ!.
Observed pattern %pportunities
Short-run Mobility levels, functional organizatio
- ' - - - - - - - - - - - - 1 and spatial distribution of activities,
Feedback at an individual level transport and land-uses, environment
(learning)
(4) The last question relates to the representation of accessibility and its meas-
urement. Experience so far in the definition of accessibility measures suggests
that an approach to their formulation should (see IRES 1995, Handy and
Niemeier 1997):
References
Amin, A. (ed.) 1994. Post-Fordism. Oxford: Blaekwell.
Andersson, A.E., Batten. D.F., Kobayashi. K. and Yoshikawa, K. (eds.) 1993. The Cosmo-
Creative Society, Berlin: Springer.
Batten, D.F. 1995. Network cities: Creative urban agglomeration for the '1"1 Century. Ur-
ban Studies 32 (2) 313-27.
Batty, M. 1995. Cities and eomp1exity. Implieations for modeling sustainability. In
Brotehie, J. Batty, M., Blakely, E., Hall, P. and Newton, P. (eds.) Cities in Competition.
Productive and Sustainable Cities tor the 21 st Century. Melboume: Longman Australia,
469-86.
Batty, M. and Xie, Y. 1996. Possible urban automata. In Besussi, E. and Ceeehini, A. (eds.)
Artificial Words and Urban Studies, Venezia: Daest, Convegni, no.l, 176-205.
Ben-Akiva, M. and Lerman, R. 1979. Disaggregate travel and mobility: Choice models and
easures of aeeessibility. In Hensher, D.A. and Stopher, P.R. (eds.) Behavioural Travel
Modeling, London: Croom Helm, 654-79.
Bennett, R.J. and MeCoshan, A. 1993. Enterprise and Human Resource Development. Liv-
erpool: Chapman.
Bertuglia, C.S., Lombardo, S. and Oeeelli, S. 1995. The interaeting ehoiee proeesses of
innovation, loeation and mobility: A compartmental approach. In Bertuglia, C.S.,
Fischer, M.M. and Preto, G. (eds.) Technological Change, Economic Development and
Space, Berlin: Springer, 118-44.
Bertuglia, C.S., and Occelli, S. 1995. Transportation, communications and patterns ofloca-
tion. In Bertuglia, C.S., Fischer, M.M. and Preto, G. (eds.) Technological Change, Eco-
nomic Development and Space, Berlin: Springer, 92-117.
Bertuglia, C.S., and Occelli, S. 1997. The impact ofthe new communications technologies
on economic-spatial systems. An agenda for future research, Paper presented at the Inter-
Revisiting the Concept of Accessibility 299
Handy, S.L. and Niemeier, D.A. 1997. Measuring accessibility: an exploration ofissues and
alternatives. Environment and Planning A 29: 1175-94.
Hansen, W.G. 1959 How accessibility shapes land use. Journal o/the American institute 0/
Planners 25:73-6.
Hanson, S. 1984. Environmental cognition and travel behaviour. In Herbert, D.T. and
Jonhston, R.J. (eds.) Geography and the Urban Environment, New York: Wiley, vol. VI,
95-126.
Hanson, S. and Schwab, M. 1987. Accessibility and intraurban travel, Environment and
Planning A 19:735-48.
Helling, A 1998. Changing intra-metropolitan accessibility in the U.S.: Evidence from
Atlanta, Diamond, D. and Massam, B.H. (eds.) Progress in Planning, London: Perga-
mon, 55-107.
Holly, B.P. 1978. The problem ofscale in time-space research. In Carlstein, T., Parkes, D.
and Thrift, N. (eds.) Time and Regional Dynamics. London: Edward Arnold, 5-18.
Ingram, R.D. 1971. The concept ofaccessibility, Regional Studies 5:101-07.
IRES 1995. Un'analisi dell'accessibilita in Piemonte. Studio di supporto aHa valutazione
delle politiche deI piano regionale dei trasporti. Occelli, S.and Gallino, T. (eds.) Turin:
Quaderni di Ricerca Ires, 74.
Knight, R.V. 1995. Knowledge-based development: policy and planning implications for
cities, Urban Studies 32:225-60.
Kobayashi, K., Sunao, S. and.Yoshikawa, K. 1993. Spatial equilibria ofknowledge produc-
tion with 'meeting-facilities.' In.Andersson, AE., Batten, D.F., Kobayashi K. and Yoshi-
kawa, K. (eds.) The Cosmo-Creative Society, Berlin: Springer, 219-44.
Koenig, J.G. 1980. Indicators of urban accessibility: Theory and application. Transporta-
tion 9:145-72.
Kwan, M-P. 1998. Space-time and integral measures of individual accessibility: a compara-
tive analysis using a a point-based framework. Geographical Analysis 30(3):191-216.
Leonardi, G. 1979. Introduzione alla teoria dell'accessibilita, Sistemi Urbani 1:65-88.
Martellato, D. 1993. Reti di interazione e tecnologia della informazione. Un'analisi dei loro
effetti sulla locaHzzazione. In Lombardo, S. and Preto, G. (eds.) innovazione e tras/or-
mazioni della cittil. Teorie, metodi e programm i per il mutamento, Milano: Collana AiS-
Rre, AngeH, 130-47.
Miles, I., and Robins, K. 1992. Making sense of information, in Robins, K. (ed.) Under-
standing information, Business, Technology and Geography, London: Belhaven, 1-26.
Morris, J.M., Dumble, P.L. and Wigan, AR. 1979. Accessibility indicators for transport
planning. Transportation Research 13A:91-109.
Papageorgiou, G. 1987. Spatial public goods, 1: Theory, Environment and Planning A
19:471-92.
Occelli, S. 1998a. Accessibility and time use in a post-Fordist urban system. Some notes for
a research agenda, Paper presented at The International Conference on Time Use, Lune-
berg, 22-25April.
OcceIli, S. 1998b. A New Perspective for methodologies in spatial and urban analysis. In
Bertuglia, C.S., Bianchi, G. and Mela, A (eds.) The City and its Sciences, Berlin:
Springer 851-72.
Pirie, G.H. 1979. Measuring accessibility: A review and Proposal, Environment and Plan-
ning A 11 :299-312.
Revisiting the Concept of Accessibility 301
Rabino, G.A. 1996. Complessita, scienza della complessita e modello dei trasporti, Le
Strade.
Rabino, G.A. and Occelli, S. 1997. Re-thinking urban system modeling: New features and a
proposal, Paper presented at the ICCS Conference, Cortona, 22-25 September.
Toulemonde, J. 1995. Should evaluation be freed from its causallinks, Evaluation in Pro-
gram Planning 1(1): 179-90.
Turoff, M. (1997) Virtuality, Communications 40(9):38-43.
Vickerman, R.W. 1974. Accessibility attraction and potential: A review of some concepts
and their use in determining mobility, Environment and Planning A 6:675-91.
Wachs, M.and Koenig, lG. 1979. Behavioral modeling, accessibility, mobility. In Hensher,
D.A. and Stopher, P.R. (eds.) Behavioural Travel Modeling. London: Croom Helm, 698-
710.
Weibull, J.W. 1980. On the numerical measurement of accessibility. Environment and
Planning A 12:53-67.
Wilson, A.G. 1971. A family of spatial interaction models and associated developments.
Environment and Planning A 3:1-32.
18 Legal Aeeess to Geographie Information:
Measuring Losses or Developing Responses?
Harlan J. Onsrud
18.1 Introduction
looking forward to more meaningful dialogue among each other and with gov-
ernment in future electronic environments, they should also be aware that past
gains made in the ability to access and build upon the works of others and gains
made towards increasing the transparency of government operations are being
eroded. Measuring and modeling the who, what, when, and where of such losses
isn't as important as understanding the manner in which such losses are occurring
and exploring alternatives by which such losses might be avoided.
18.2 Background
The United States is unique in the world in the broad access to information that its
laws support. Areas of the law influencing access to information, geographic or
otherwise, include intellectual property law, freedom of information law, privacy
law, electronic contracting law, and anti trust law, as weil as several other areas of
the law. Two generalizations about the interoperation of these laws appear ger-
mane to the topic of this book.
The First Generalization. The forms that these laws take in the United States
allow greater access to government iriformation at the local, state, and national
government levels and use of that information than is generally allowed in other
nations. For instance, few nations have national freedom of information laws that
allow citizens broad general access to the public records of government. 1 Even in
those nations that do allow such access, citizens are not allowed typically to add
value to such information and resell it without the permission of government as
they are allowed to do in the United States. In addition, the United States goes
much further than other nations since it actually imposes affirmative obligations
on federal agencies to actively disseminate their information as defined by the
provisions of OMB Circular A-130 (June 1993). Agencies are particularly encour-
aged to disseminate raw content upon which value-added products may be built by
the private sector and to do so at the cost of dissemination, with no imposition of
restrictions on the use of the data and through a diversity of channels. The core
provisions of OMB Circular A-130 were incorporated into the Paperwork Reduc-
tion Act of 1995 (PRA) and that act additionaily encourages the use of information
technologies by agencies for providing public access, rather than relying on cum-
bersome Freedom of Information Act (FOIA) processes. With the expanded use of
World Wide Web servers by federal agencies, the cost of dissemination for many
1 Fifteen nations that have general open government records laws, several of which are
recent, are listed at http://www.cfoi.org.uk/foioverseas.html.
Legal Aeeess to Geographie Information 305
federal governrnent data sets has beeome negligible and, thus, these data sets are
now freely available to anyone with the ability to access them over the Internet. 2
Actions have also been taken at the federal level specifically related to spatial
information and agency contributions to building the National Spatial Data Infra-
structure (NSDI). The Office of Management and Budget (OMB) established the
Federal Geographie Data Committee (FGDC) in its 1990 revision of Circular A-
16, Coordination of Surveying, Mapping, and Related Spatial Data Activities.
FGDC is now composed of representatives from 17 Cabinet level and independent
Federal agencies. In April 1994, President Clinton signed Executive Order 12906
that called for the establishment of a coordinated National Spatial Data Infrastruc-
ture (NSDI) as part of the evolving National Information Infrastructure (NIl), and
FGDC was charged with coordinating the federal governrnent's development of
the NSDI. In this executive order, FGDC was given a mandate to involve state,
local and tribai governrnents, academia and the private sector in coordinating the
development of the NSDI. The roles of various parties and their relationships in
moving towards a common NSDI vision are being developed over time.
Similar to the federal situation, open access laws exist in most of our states that
impose similar broad principles of access by citizens to the records of state and
local governrnent agencies. Because of this atmosphere of openness, many local
municipal, county, and state governrnents have voluntarily been making geo-
graphie data sets available on the web for general use by for-profit businesses, not-
for-profit organizations, and citizens generally. 3
Rights of access to governrnent information and the atmosphere of openness that
these rights engender are currently very significant. We should not allow these
rights to be chipped away at through growing numbers of legislative exceptions
for geographie information databases and other governrnent databases.
The Second Generalization. u.s. law grants individuals greater leeway to use
the work products of others without perm iss ion than is typically granted by the
laws of other nations. The law of copyright in the United States grants fewer own-
ership interests in intellectual works and greater access to the work products of
others than in perhaps any other industrialized nation. Again, we should not aban-
2 Data sets available from U.S. federal ageneies may be traeed typieally through their offi-
cial web sites indexed at http://leweb.1oe.gov/globallexeeutive/fed.html. Examples of spa-
tial data sets available from federal ageneies inc1ude those found through
http://www.usgs.gov/themes/info.html, http://www.eensus.gov/ftp/pub/www/tiger/,
http://www.epa.gov/epahome/data, and http://gemd.gsfe.nasa.gov/.
3 See http://reeorder.marieopa.gov/reeorder/imagingl for an example of open web aeeess to
deeds and plats; see http://www.ei.ontario.ea.us/for an example of aceess to eommunity
geographie information; or see http://fgdec1earhs.er.usgs.gov/ for state and eommunity
c1earinghouse nodes.
306 H. J. Onsrud
don this high level of access to the information sources and products produced by
others without substantial social benefit reasons for doing so.
The United States has productive scientific and commercial database communi-
ties that are the envy of the world. The vitality of these two communities go hand-
in-hand. The govemment, commercial, and not-for-profit sectors have all bene-
fited by the past balance of legal policies that has minimized the need to pay for
access and use of data drawn from other govemment, commercial and not-for-
profit sector sources. For-profit private sector creators of derivative databases
have gained as much or more from this policy balance than any other societal sec-
tor. Although both new innovations and past investments are protected under U.S.
law, the current complex balance in the law provides a tension in favor of new
innovations over past investments. This keeps competition and the need to inno-
vate high. The level of activity and growth of the database industry in the United
States in comparison to other nations is evidence that a flexible ability to draw
from and build upon the data sets collected by others is highly desirable for eco-
nomic development. The strong health and high activity of the U.S. database in-
dustry also suggests that there is no crisis at hand that would warrant granting the
private commercial sector immediate greater control over the data it collects.
Geographic data, like many other forms of scientific and technical data, possess
the classic characteristics of public goods. That is, geographic data are typically
nonrival and nonexcludable. In short, a nonrival good is one that may be con-
sumed without detracting in the least from consumption of the same good by oth-
ers. By example, the use of digital data for finding one's way does not make the
same data any less useful to others for finding their way. A nonexcludable good is
one whose benefits are available to all once the good is provided. Once nonex-
cludable benefits are provided they may be very difficult or perhaps impossible to
exclude from others even though others may not have helped pay for the good.
Due to the ability to copy data electronically and transfer it over networks at neg-
ligible cost, data made available to small numbers of persons are often transferred
to much wider audiences regardless of the contractual provisions stipulated and
the technical protection techniques used, such as electronic watermarking or en-
cryption. Geographic data, as wen as many other forms of factual data, are addi-
tionally nonexcludable in that others may recollect the same or substitutable data.
It is because ofthese public goods characteristics that much geographic data in the
United States has been collected in the past by govemment and access provided to
all as a general public benefit.
The commercial sector has little incentive to collect data possessing the charac-
teristics of a public good unless some form of subsidy is provided. Copyright law,
by example, is a subsidy established by law that provides an incentive for creators
to select and arrange data to make it more useful for society. Any consumer of the
original and creative selection or arrangement is required to subsidize the creator
(or owner) by paying a higher price than would otherwise be required. Because
education and research also have socially desirable outcomes, those copying an
authored data set for education and research activities are not required to pay the
Legal Aeeess to Geographie Information 307
subsidy under certain circumstances (i.e., see the fair use provisions of the Us.
Copyright Act). Similarly, any other person or concem may extract factual infor-
mation from a data set without paying a subsidy or asking for permission as long
as that person doesn't copy creative original aspects ofthe work and as long as that
person is not breaching a contract with the owner. Further, the first safe doctrine
of copyright law (17 U.S.c., section 109) specifically authorizes the owner of a
legally acquired copy to seil or otherwise dispose ofthe copy and thus subsequent
sales of a copy do not require further subsidy payments to the creator (e.g., resale
of a typically purchased book or CD). These examples iIIustrate that copyright
owners have never had the ability to extract the full market value from their
works. Protection for copyright owners has been established by the law over time
at a level that provides strong incentives to produce works of authorship but not at
a level so high that author rights significantly impinge upon society's interest in
allowing the public to make reasonable use of authored works. Society benefits
more in terms of advancement of science, the arts, and the overall economy if rea-
sonable leeway is provided to allow each of us to draw from and build upon the
works of others.
It is noteworthy that the United States gained significant economic strength and
dominance in information technologies, as weil as in research and technology
generally, at a time when its information laws were very different from those of
other nations. The role of U.S. laws and policies in supporting an open environ-
ment of access to scientific data for the commercial and science sectors and the
role of U.S. laws in ensuring access to govemment data sets should not be over-
looked when exploring the competitive success of U.S. businesses and scientists.
Observation of recent actions in Congress and legislative actions at the state and
local govemment levels suggest that the nation is back tracking on its openness
principles rather than extending them. Some observations of recent lawmaking in
action include the following.
firms that can afford the data benefits primarily those privileged firms at the ex-
pense of the general public and the loss of widespread general benefits to the
community. Those who seek to impose restrictions on citizen access should be
required to overcome the underlying policy arguments on which such laws are
based, foremost of which are that open access keeps government accountable and
that open access to government information has far greater long-term economic
benefits for a community than does pursuing revenue generation approaches.
(3) Extension of Time for Copyright. The special genius of the Vnited States
copyright system has been its emphasis on an appropriate balance of public and
private interests. V.S. dominance in international trade in current products of au-
thorship has been made possible because of the rich and vibrant public domain
passed down from earlier authors. Proposed legislation (H.R. 989) would extend
the term of copyright protection for all copyrights, including copyrights on exist-
ing works, by 20 years. For individual authors, the copyright term would extend
Legal Access to Geographie Information 309
for 70 years after the death of the author, while corporate authors would have a
term of protection of 95 years. Unpublished or anonymous works would be pro-
tected for aperiod of 120 years after their creation. The enactment of this legisla-
tion would impose substantial costs on the U.S. general public without supplying
any public benefit. It would provide a windfall to the heirs and assignees of au-
thors long since deceased, at the expense of the general public, and impair the
ability of living authors to build on the cultural legacy of the past. The proposed
extension would supply no additional incentive to the creation of new works - and
it obviously supplies no incentive to the creation of works already in existence.
The notion that copyright is supposed to be a welfare system to two generations of
descendants has never been a part of American copyright philosophy. It is not
wifair that a work enters the public domain 50 years after the death of its author.
Rather, that is an integral part of the social bargain on which our highly successful
system has always been based. After supplying a royalty stream for such a long
time, these old works should be available as bases on which current authors can
continue to create culturally and economically valuable new products (extracted
and rearranged from Karjala (1995)).
NOTE: In spite of such arguments and widespread opposition by the academic and li-
brary communities, both the V.S. Senate and House passed on 7 October 1998 a 20-
year extension ofthe then existing life-plus-50-year copyright term4 .
NOTE: The proposed article to the Uniform Cornmereial Code has been withdrawn
but now is being pursued by the Uniform Law Cornmissioners as an independent aet
under the title Uniform Computer Information Transactions Acr. Thus the proposal is
still under aetive eonsideration even though widely opposed by very diverse groups
ranging from film studios to eonsumer groups.
These are but a few illustrative examples of major attempts continually occur-
ring in our legislative halls in attempts to restrict citizen access to public domain
and government information. Thus, the assumptions of access to which we may
have become accustomed when we operated in a paper world should not be taken
for granted as data, information, and works of knowledge are transferred more and
more by electronic means.
5 See http://www.2BGuide.eom.
Legal Aeeess to Geographie Information 311
ing the trend of building walls around government information would be to require
of government officials an 'information access impact statement' for any major
state or local government action significantly affecting the quality of citizen ac-
cess to government information. Many additional political and legislative lessons
may be gleaned from the experiences of the environmental community in protect-
ing the environmental commons.
Looking outside of the environmental-Iaw realm, additional specific legal provi-
sions or policies that might be advocated in expanding information access to sci-
entific and technical data in the public interest might include the following:
(a) Legislate the first safe doctrine in networked electronic environments in in-
stances where technology allows no more than one user of a purchased intel-
lectual work at a time.
DiscussioD. The library community has been constructing its own set of
standard licensing provisions with the implication that many librarians and
the economic bloc they represent will no longer contract or license with elec-
tronic publishers that do not adhere to the library community's recommended
licensing provisions6 • Is this approach realistic? How may the library com-
munity position be strengthened without running afoul of antitrust laws and
without devoting larger proportions of the library community's resources
over time to licensing negotiations and the tracking of adherence to licensing
provisions? Although the library as an institution may try to act in the best
interests of its user community, in an economic marketplace negotiation en-
vironment, when push comes to shove, will there be a strong tendency for the
library to act in its own best economic interests, and to sacrifice some social
welfare interests of its user community?
Discussion. Under this approach, virtually all copyright or other legal rights
in authored works or data sets might be transferred by professors to publish-
ers but, with the exception of the right of first publication, transfer of exclu-
sive rights would not be permitted. Retention of full non-exclusive rights by
universities would be for the purpose of letting the works or data enter the
public domain at the option ofthe university, particularly under marketplace-
failure circumstances. Because transfer of exclusive rights to private parties
works against development of a common public domain in scientific works
and technical data sets and thus works against scientific advancement, those
professors and researchers transferring exclusive rights to publishers or oth-
ers would have their works devalued by the university recognition and re-
ward system processes. What would such a model policy for universities or
funding agencies look like?
I view all of the approaches mentioned so far as pragmatic proposals that may be
used internal to the current legal system to make our society more responsive to
protecting public access and the public commons in information. Within the envi-
ronmental realm, another approach that has been far less successful to date has
been to argue that there is something inherently wrong and unjust about the whole
concept of real property ownership and to move to different models and concepts
of ownership and rights in land. To deal with the environmental problems of the
nation and the world, the argument is made that the current legal system can't
support an appropriate solution and therefore we need to step outside of the con-
straints of the current legal and political system to arrive at systems that would be
more responsive to the needs ofthe environment.
Vandana Shiva suggests that there is something rotten at the core of ownership
claims in information and treatment of information as a commodity to be sought
and sold. Rather than hone the existing legal and social models that assume that
the current inequities in society are a given, she argues that there is a need to ex-
plore whole new models and theories of rights to access and use of data, informa-
tion, and knowledge works. She offers a non-western, global, and community con-
trol perspective in which neither the state nor the market provide the organizing
principles of how people live and how nature's wealth is owned and used (Shiva
1994, 1997).
Practical incremental approaches in expanding rights to information within the
existing legal framework and wholesale reevaluation approaches are not mutually
exclusive. Even though treatment of information as a commodity may be a social
construct, information is, in fact, being treated as a commodity in our local com-
munities, at the national level and throughout the globe on a day-to-day commer-
cial basis as weIl as through the imposition of intellectual property rights laws.
One may try to limit and adapt intellectual property rights laws to help ensure con-
tinued access to information and the continued development of public domain data
or one may suggest complete new models or views on how control over informa-
tion should be handled. The academic community should pursue both of these
approaches.
As we already know, access, whether in the form of technically meaningful ac-
cess or legal access does not guarantee power. Nor do consensus building or other
participatory processes equalize political power. Measuring or modeling access, as
weIl, does very little to affect power relationships. However, all of these are useful
tools for aiding in struggles to gain political power. Paul Schroeder has noted that
changing the conditions of power in society would have a substantial influence on
changing the conditions of access to and handling of information (Schroeder
1998). Therefore, the implication is that to change the conditions of access one
should focus primarily on altering power bases in society. Yet, the alternative ap-
proach of direcdy changing specific rights in information also has an influence
over power and wealth in society. If this were not the case we would not see such
intense lobbying in Congress over rights in information at the current time. Near
the core of this power struggle lies the debate over development of a concept of
314 H. J. Onsrud
human rights that could counteract or limit existing corporate and govemment
agency powers in US. society.
18.6 Conclusion
In the past, U.S. law has supported the proposition that citizens should have broad
and open access to government information at local, state, and national govern-
ment levels. In addition, U.S. laws have granted greater leeway than the laws of
other nations to use the work products of others without permission in order that
access and new innovations should be promoted and take precedence over wealth
generated from old innovations. Both of these general principles are being se-
verely challenged as publishers and government agencies use the threat of digital
technology as an opportunity to limit the rights of citizens to access information.
Assuming that legal access to government information and other forms of infor-
mation and intellectual works may be maintained, most of us want increased ac-
cess to information as more and more of our daily activities are accomplished and
relationships are established within digital environments. We want technical ac-
cess to data, information or knowledge that is efficient, effective, and responsive
to our specific needs. We want pro ce dural capabilities and methods that will allow
groups affected by decisions to be engaged with each other in constructive dia-
logue. We want access that is timely and understandable so that interested groups
may constructively participate with government in more democratic decision mak-
ing.
Although rights of access to information are insufficient conditions in them-
selves to achieve these goals, they are necessary and critical conditions. Providing
and protecting legal access to information and knowledge works is at least as im-
portant as expanding effective and efficient technical access or developing means
for measuring access. In addition, while each of these may be necessary and
highly constructive societal activities, substantial gains in power for citizens and
citizens groups will require new approaches to ownership (e.g., as suggested by
Shiva) or major realignments in our existing constitutional framework (e.g., as
suggested by Black).
Acknowledgments
Material for this ehapter was developed in preparation for two specialist meetings funded
by the Varenius projeet of NCGIA; Empowerment, Marginalization, and Publie Partieipa-
tion GIS, NCGIA Varenius Specialist Meeting, Santa Barbara CA, Oet. 1998 and Measur-
ing and Representing Aeeessibility in the Information Age, NCGIA Varenius Special ist
Meeting, Asilomar, Paeifie Grove CA, Nov. 1998.
316 H. 1. Onsrud
References
Black, C.L., Jr. 1997. A New Birth ofFreedom: Human Rights, Named and Unnamed. New
Y ork: GrossetIPutnarn.
Guernsey, L. 1998. A provost challenges his faculty to retain copyright in articles, Chroni-
cle of Higher Education, 18 Sept.
Karjala, D. 1995. WRlTTEN TESTIMONY ofDENNIS S. KARJALA, Professor ofLaw,
Arizona State University representing UNITED STATES COPYRIGHT AND
lNTELLECTUAL PROPERTY LA W PROFESSORS before HOUSE OF
REPRESENTATIVES COMMITTEE ON THE JUDICIARY SUBCOMMITTEE ON
COURTS AND INTELLECTUAL PROPERTY on H.R. 989. A Bill to Amend Title 17,
United States Code With Respect to the Duration of Copyright, and for Other Purposes.
Rayburn House Office Building, Washington DC, l3 July.
Lober, D.J. 1995. Resolving the siting impasse: modeling social and environmental loca-
tional criteria with a Geographie Iinformation System, American Planning Association
(APA) Journal, Autumn, 482-95.
Nunberg, G. 1998. Willlibraries survive? The American Prospect Nov/Dec: 16-23.
Nyerges, T., and Jankowski. P. 1998. Empirical research strategies for investigating the use
of public participation GIS, international workshop on groupware for urban planning,
Universite C. Bernard Lyon I. Lyon, France: 3-4 Feb.
Onsrud, HJ. 1998. Tragedy of the Information Commons. In Taylor, D.R.F. (ed.) Policy
Issues in Cartography, Pergarnon: Elsevier Science Ud., 141-58.
Onsrud, H.J., Johnson, J.P., and Winnecki, J. 1996. GIS dissemination policy: two surveys
and a suggested approach, Journal ofUrban and Regional Information Systems 8:8-23.
Sarnuelson, P. 1996. The copyright grab. Wired Archive 4.01 (lan).
http://www.wired.comlwired/archive/4.01/white.paper.-pr.htm
Samuelson, P. 1998, E-mail to Cyberia-L, 16 July.
Schroeder, P. 1998. Statement made during the NCGIA Varenius Specialist Meeting on
Empowerment, Marginalization, and Public Participation GIS, Santa Barbara,CA: Oct.
Shiva, V. 1994.The recovery ofthe commons, Alternative Radio, Colorado College, Colo-
rado Springs, PO Box 551, Boulder, CO 80306, 24 Sept.
Shiva, V. 1997. Biopiracy: The Plunder of Nature and Knowledge. Boston: South End
Press.
Stefik, M. 1997. Trusted systems, Scientific American 276(3):78-81.
Stefik, M., and Lavendel, G. 1997. Libraries and Digital Property Rights. In C. Peters and
C. Thanos (eds.) Research and Advanced Technology for Digital Libraries, Proceedings
ofthe First European Conference, ECDL '97, Pisa, Italy, 1-3 September 1997. Berlin:
Springer.
19 Qualitative GIS: To Mediate, Not Dominate
Robert Mugerauer
Barbara Parmenter and I are conducting aseries of projects to clarity logical and
pragmatic alternatives. We begin with two assumptions: (1) GIS is structured on
formal Euclidean geometry for spatial representation and on alpha-numeric data-
base principles for informational content; and (2) current data bases represent
Newtonian-Cartesian spatial conceptions and practices. What follows is our criti-
cal question: Given these two descriptively defining characteristics, is it the case,
either theoretically and/or practically, that GIS must operate on these Euclidean-
Newtonian-Cartesian principles only? If GIS is not so limited, then Qualitative
GIS could be constructed on non-Newtonian, non-Cartesian, perhaps non-
Euclidean databases - which can be found in or derived from the already existing,
extensive ethnographic research literature and other existing data sources. On the
other hand, if GIS is strictly contained within Euclidean-Newtonian princip\es of
organization, then Qualitative GIS, strictly speaking, is impossible. The best that
could be accomplished would be a translation of qualitative properties into
Euclidean and alphanumeric representations. Even here, however, we have the
possibility of two kinds of qualitative GIS. One such qualitative GIS would com-
plement current GIS by inserting or encoding various kinds of hyper-media into
standard GIS bases, actually superimposing qualitatively distinct information upon
that standard base. The end result would be a kind of updated medieval, multi-
perceptual mapping. Recall how medieval mappings regularly presented naviga-
tion information, along with glosses and drawings that surrounded or overwrote
the basic cartography with story-telling, imaginative, theological, and other modes
of information. A contemporary version of this would electronically insert per-
sonal, local, and imaginative narrations, images, and other perceptual-qualitative
information over or through the standard GIS spatial layout. Alternately, it is pos-
sible to model mathematically various spatial configurations, for example, to rep-
resent qualitatively differentiated spatializations (raising the issue of whether such
a format would be a mapping or a modeling, a question that does not need to be
settled here). In either of the last two cases, though we would not have 'Qualita-
tive' GIS strictly speaking, we nonetheless would have something close enough to
it that, for non-specialized purposes, we would not have to apologize for and could
drop the quotation marks, setting it off more rigorously according to its epistemo-
logical grounding?
With either ofthese qualitative modes, the result would be a GIS that presents a
set of alternative geographies and alternative ways of visualizing those spaces and
2 I want to thank Professor Parmenter for her valuable contributions to this project. Not
only did she keep me on the straight and narrow by providing normative control for cor-
rect use of concepts and technical terms, and provide helpful critique on the early drafts
of this paper, but she continues to show a wonderful openness to theoretical and practical
exploration ofthe topic. As I say to our students, we make a good team, since she knows
what GIS actuaUy is, while I, relatively unencumbered by facts, then can safely propose
wild-eyed ideas. See at http://mather.ar.utexas.edu/students/cadlab/spicewoodl for more
information on our current attempt at doing Qualitative GIS for a grass-roots, neighbor-
hood natural and settlement environments project. Ifyou have questions about the project,
contact Barbara [parmentr@uts.cc.utexas.edu] or me [drbob@mail. utexas.edu]. I discuss
the project briefly at the end of this paper.
Qualitative GIS 319
19.2 Problems
On whose behalf do GIS technical specialists gather and speak? It would seem
presumptuous to say, since those who may be interested in more access to GIS
would need to speak for themselves. But, they may not come forward unless in-
vited, unless encouraged. So, we need to formulate an invitation, that is, to begin
to create an opening in which they would be weIcome and in which there would
be a point to their coming. Of course, we are responding to the well-documented
need that exists because of a gulf between those who have and use electronic tele-
communication-information technology and those who do not. 3
Researchers, practitioners, and policy makers recognize increasingly that the
emerging division between 'haves' and 'have-nots' is no longer between geo-
graphically distinct first and third or fourth worlds divided according to degree of
modernization-industrialization. Instead, it is between groups with and without
access to the new information technologies and the power these bring. These new
'dual societies' often are found side by side, in the same cities and regions, in
Washington, OC and Mexico City, in rural California and France (CasteIls, Mol-
lenkkopf, and Robson 1998, Sanyal 1996).
Because highly developed information technologies, such as GIS, are both a
product of and a me ans to develop our scientific and capital-intensive culture, we
rightly assume that there are problems when these technologies are not as wide-
spread as they might be. However, the motives and reasons of various constituen-
cies using, not using, and promoting the spread of information technologies are
varied.
3 The evidence is presented in all formats and for many audiences: in books by and for aca-
demics and researchers (Investors Business Daily 1998, Loader 1998, NTIA 1997, Schil-
ler 1996). There is a rapidly growing literature on the subject. Feenberg and Hannay
(1995) is theoretically noteworthy.
320 R. Mugerauer
tion of not only spatial elements, but economic, cultural, and any other kind of data we
wish. Given the ability to see the way different dimensions of the world do or do not
correlate, we can proceed with planning ways to change conditions to more cIosely re-
alize the model we ultimately des ire.
Among the assumptions, several concern the character of space. Though New-
tonian physics has been supplanted by relativity, Newton's own theory of abstract
and relative spaces remains pragmatically adequate to explain and operate within
our earthly geo-political realms. Thus, we still use variations of Euclidean geome-
try and the concepts of absolute, abstract space in GIS because they are operation-
ally correct and adequate. The definition of a 'good' map is one that corresponds
correctly, point-by-point, with the features of the earth that objectively exist. This
uses the Newtonian idea that in order for there to be a material world at all and for
it to operate with law-like movements and forces, there first must be a containing
envelope of space. This absolutely existing space (independent of and prior to the
material bodies that come to occupy part of it) is not directly experienced, but ab-
stractly understood through the mathematical sciences. Because this space is inde-
pendent of material bodies, and a condition for their appearance, it itself is homo-
geneous--the same throughout. Differences within space are accounted for in
terms of bodies and forces among bodies. Congruently, this homogeneous space is
isotropie; that is, no 'direction' is inherently different than any other, much less
privileged. Directional differences are purely a matter of our humanly oriented
experiential-relational space.
Correlated with these spatial conceptions, developed over hundreds of years and
displacing earlier Greek-based theories of relative, heterogeneous, and anisotropie
space, there are epistemological assumptions of positive science. In brief, these
include the following: (a) It is held that the world consists of at least space, physi-
cal materials, and forces of relation and change among the elements of matter. (b)
The human mind (and parallellinguistic and symbolic systems) has the capacity to
re-present objective states of affairs in our thought processes and symbolic repre-
sentations. Thus, (c) what is true is what is a correct representation. Correct con-
ceptual representations are held to work best (or perhaps only) in tight logically
univocal concepts and in mathematics. In our area of concern - the visualization
of data sets - the good/true map is one with features that correctly (completely and
consistently) correspond to and re-present the topographical state-of-affairs. Simi-
larly, data sets (e.g., the location of power and utility lines, land valuation and tax
figures, zoning information, etc.) are true when they correspond correctly to the
physical or social phenomena they represent and also fit correct1y to the map it-
self. GIS then consists of a 'nested' series of representations that have their value
in being correct and manipulatable re-presentations of objective states of affairs.
But, reality is not so simple. Certainly this is not the place to 'refute' or 'amend'
the above assumptions. Here I can only assert that they are 'correct,' but incom-
plete and historicaIly-politicaIly constituted; that is, not straightforwardly anything
like the whole and entire truth. This alternative position seems weIl established by
current theoretical debates in the history of science, hermeneutics, and critical
theory, to which we can refer should we wish (see Lefebvre 1991, Mugerauer
322 R. Mugerauer
1991, D'Amico 1989, Gadamer 1989, McIntyre 1988, Heelan 1983, and Heideg-
ger 1977). Suffice it to say that 'facts' or 'data' are not self-selective or self-
validating; what becomes a fact or data-point or counts as 'information' does so
only within the context of a conceptual-practical system, which itself has a histori-
cal, cultural context of limitations and aspirations, insights and blindness, fears
and hopes. In short, what counts as information or even as a geographie feature is
a eoneeptual-pragmatie representation that results from diseernment, seleetion,
and suppression among alternatives within a historieal, eultural world system.
In addition, the lifeworld experience of places is primary and the conceptual
constitution and grasp of abstract space a secondary and derivative development.
As case studies in phenomenology, ethnology, and psychology demonstrate, in our
experience, places appear as heterogeneous (not homogeneous), as a function of
relationships to other people, places, and things (that is, relative, not absolute), and
with directional differences of up and down, back and front, right and left, all of
which are physiologically, psychologically, and symbolically charged (not iso-
tropiC).4 Thus, our lived geographical experiences display features exactly the
opposite of those attributed to space by the reigning conceptions of GIS. Since the
similarities and differences of places experienced among individuals, groups, and
entire cultures are among the chief sources of social cooperation and conflict, and
of the opportunities and obstacles that we seek to consider, it is not politically suf-
ficient or proper to operate from the limited conceptions of dominant positive sc i-
ence.
That there are alternative geographies and alternative ways of visualizing spaces
and places is patently obvious. One wonderful advantage of GIS is that it presents
its diverse data visually. This is positive because, at some levels at least, those
who are not fluent with concepts or numbers can interact with visual information -
though social conventions are an enormous factor in sharing or mediating between
'creators' and 'users.' Further, cultural history shows that while some instruments
are highly directive or limiting to those who use and interpret them, there are
minimallimitations with simple drawing instruments (sticks scratching maps in
the sand or dirt; drawing on hide, bark, paper, and stone with pencils or powdered-
colored pigments; weaving various materials, or in the oral vers ions of mapping
that specify places and routes with song and story. As to the latter, Bruce Chatwin
(1987) nicely presents the Australian aboriginal tradition in which a physical-
spiritual world is mapped by stories and songs; Inuit and other native American
4 Within the large body ofwork in phenomenology and Gestalt psychology, of special note
is the work ofRudolf Amheim (1986), Thomas Thiis-Evensen (1984), and Maurice Mer-
leau-Ponty (1979).
Qualitative GIS 323
traditions of oral and visual mapping are eovered in Bravo (1996), Rundstrom
(1995), Moodie (1994), Brody (1988, 1989), Turnbull (1989), Aberley and Lewis
(1998), W oodward and Lewis (1998), Walhus (1977), and Hisatake (1986).
Even a brief look at a variety of mappings makes clear that a range of eompel-
ling lifeworld geographies, rieh in understanding, interpretations, and information,
is brought forth by the designs and sayings/namings of many peoples where the
visual and verbal systems are artieulated in loeal or dialeetieal 'mother tongues'
(whieh eertainly are not the same as the systems ofthe univoeal eoneepts in West-
ern scienees, philosophy, and other diseursive formations). The variation in map-
ping beeomes obvious even in the simple set of figures provided in Figures 19.1-
19.5.
Figure 19.1. Standard Western cartographic representations: Mercator and Gnomic projec-
tions
Given the assumptions noted above that ground and drive GIS, it is elear how
the now-standardized forms of seientifie eartography provide the exemplars: they
are taken to be the correet representations of the objeetive state of affairs. From
this point ofview, the other modes ofmapping are interesting, perhaps, but 'ineor-
reet,' or 'deviant,' or representative of some other dimension (such as the makers'
dreams, feelings, impressions, limited pereeptions, ete.), but not of the objeetive
state ofthe world.
The assumptions diseussed here and the attitude toward the 'ineorreet' is nieely
put by Peter Gould and Rodney White (1980).5 Though they eertainly are deeent
and well-intended persons, as are the rest of my positivistie eolleagues, they ulti-
mately display, use, and promote the austere judgments of positive seienee. When
they examine 'the eorrelation between preferenees and aeeuraey of loeation' they
5 These authors, pioneers in perception and mental mapping research, are sensitive that
dass and economic resources make differences and they do describe the sociological un-
derstanding that comes from taking groups' perceptions as they are. Of course, their book
is innocent ofthe theoretical sophistications developed here.
324 R. Mugerauer
straightforwardly assume that the objective character of the terrain and correctness
of representation are what matter.
For our first, and to geographers, discouraging, plunge into spatial ignorance, we shall
examine the situation in North Dakota. University students were asked to record the
names of states on an outline map, and by recording the proportion of errors we can
draw contour lines enclosing areas of equal misidentification. (Gould and White 1980,
82-83)
..
,
~ _-"- ___
~" ,:,"""""",~
' '-'''''''' . . . .
. . . -A........
Since they are scientific, they attempt to explain the cause of distortions and
'barriers to information flow' in terms commonly used in GIS and other informa-
tion-communication studies and policies. They work to show that the number of
transmissions (direcdy related to the number ofpeople) and the degree offamiliar-
ity (direcdy related to proximity to geographical features described) explain the
degree of accuracy or distortion of representations.
After we have taken the logarithms of information, population, and distance from [our
research areal, we can write: log information = -l.38 + 0.87 log population - 0.40 log
distance .... There is a very strong and significant relationship of information to both
these predictive variables. (1980, 93)
The quiet force behind what Gould and White say lies in its comprehensive
grasp of our increasingly global economic, legal, political, military, inteUectual,
and other institutions. Being able to present one's case in logical, linear terms, with
quantitative evidence, is essential if one is to obtain a grant, be hired or promoted
in the realms of research and technology, or convince a jury, city council, or gov-
emment agency to grant one's request. This is why, no matter what one's episte-
mological or political position, it is critical for those inside and outside the domi-
nant realm to learn standard GIS. Since the standard view is what exercises power
in the world today, and increasingly so, one has to be able to understand it and
participate in it oi become excluded from power of all sorts. To argue against the
importance of the reigning view conceming objectively arranged space and its
technologies, including GIS, would be poindess. Thus, a first conclusion: those
who have no access to GIS need to find a way to leam it, to acquire access to it, to
use it. But, this is a minimal consequence of our reflections, for it is pragmatically
harmful if one's lack of technology leads to being eliminated from the world, ei-
ther effectively or actually.
Figure 194
. . ABI ac kfloot tiPI
.. cover, painted with war episo
. des
Qualitative GIS 327
Figure 19.5. A map from Paulo Freire's pedagogical exercises. Reprinted by permission of
Continuum International Publishing Group from Education for Critical Consciousness.
Copyright 1973 by Paulo Freire.
328 R. Mugerauer
The problems with current GIS systems and their social uses form a relatively
simple cluster, no matter what one's personal political or intellectual position.
Since the dominant technological systems are grounded upon the post-renaissance,
post-enlightenment system of rational-mathematical science, understanding the
world depends, as Galileo already noted, on ability to do mathematics. Today,
alternate symbolic systems are becoming available in forms we call 'user friendly'
but which actually amount to translation of mathematical-Iogical codes into other
representational forms, typically iconographic. Thus, while the 'driver' does not
need to write code or understand the workings 'under the hood' she does need to
have 'dashboard knowledge.' That is, the user must be literate in and dexterous at
symbol distinctions, sequencing and other analytic-Iogical relations and opera-
tions, as weIl as in certain kinds of behavior routines.
Even setting aside the enormous pedagogical and political problems of how to
help others become computer and GIS literate, there remain several bitter realities
facing policy decisions. Given that there are many people who do not have access
to the dominant GIS technologies and worldviews, there is not agreement on what
to make of this fact. Currently, those with the information technologies live in a
world where those without it largely are ignored. Apparently, many GIS special-
ists are or should be concerned with finding solutions to these problems. Most of
us apparently believe in the value of inclusion of disenfranchised groups and in
cooperation with other world systems. But, we need to be critically aware of our
diverse motives and assumptions, lest we ourselves act imperialistically. Not sur-
prisingly, even the weIl-meaning formal directives behind the NCGIA Varenius
Project seem to consider the needs and possible remedies in terms of 'concepts'
that 'reconceptualize, measure, represent, monitor, and plan for the new emergent
geographies,6, thus almost inevitably casting the project in the very terms of the
dominant 'imperialistic' educational process. This is perverse since it is precisely
by their differences from the standard and dominant categories that the already
marginalized groups constitute their identity. In addition - though unavoidably -
these 'have-nots' (the learners) are required to consciously or unconsciously con-
form by internalizing and using the very 'normative' concepts, maps, and images
of the dominating groups (the teachers, fund-providers, and ultimately the 'host'
social-conceptual-technological systems or cultures), of which more shortly. In its
current form, it would appear that weIl-intended projects such as Varenius are
reconceptualizing the issue in the same rationalistic terms that will perpetuate the
inequality of accessibility opportunities, insofar as the latter have any substantial
economic or political force, or further obliterate local, differentiated groups' iden-
tities.
6 Cited from the NCGIA Varenius Project's 'Call For Participation'in 1998.
Qualitative GIS 329
ist groups' deliberations and consequent actions to jigure out how to cooperate
with those who would affirm their own distinctive worldviews and geographical
information systems; that iso with those who do not want to lose their own identi-
ties and ways of life just because they might have the opportunity to obtain access
to ours.
The proposed partial solution to many of the problems outlined above aims at an
affirmation of the identities and differences among individuals and groups within
the context of a shared set of worlds. This would be the contemporary, informa-
tion-age version of e pluribus Unum. Without the 'one,' we have chaos - anarchy,
if not war; without the 'many,' we have totalitarianism.
The outcome envisioned here is intended to be simple and realistic. It is simple
in that, opposite to a monoculture, which seems to spell doom to human social
groups just as surely as to soi! and crops, it envisions living in a multiply-cultured,
non-isolated, set of worlds whi!e maintaining several, possibly changing, identi-
ties. The vision is realistic in that it has operated across time and space for thou-
sands of years. Very few people actually have been or have remained members of
absolutely undifferentiated monocultures. Even within small primal groups there
are multiple sub-cultures: men's and women's groups, earth and sky groups, mon-
key and snake people, children and post-initiates, gatherers, warriors, and shamans
in dynamic relationships. Even among the earliest and most c10sed groups there
are those who operate at the borders, leaming and using the languages and mate-
rial items of neighbors. The ancient trade of colored stones and weapons worked
because groups with strong focal identities nonetheless had ways to interact with
others who, in effect, lived in different worlds. The same phenomenon continues
with the millions of migrants in today's world. Think of the worldwide phenome-
non of children of immigrants mediating between the 'old world' culture of the
transplanted grandparents and the host culture ofthe streets.
Transculturation does work. How? Note, here I am not talking about replacing
one culture with another; whether freely chosen or forced. That phenomenon has
to do with the operations of monoculturation. I mean the process whereby one
maintains one's own initial cultural world and comes to participate in another, or
several others, which also become one's own, while remaining able to pass back to
inhabit, even to deepen, one's original 'horne.' Again, most ofus do this regularly,
as would be apparent if we discussed our own lives as sons and daughters and
parents, as Irish-American researchers studying Chilean economics, as academics
who also repair and race motorcyc1es, and so on.
The process of which we are speaking is one of mediation, where some people
open to each other, help each other to cross over and back, between cultures. This
Qualitative GIS 331
Lankshear, Paters, and Knobel 1996). To teach the corpus of knowledge and pro-
cedure that is the heart of any tradition, we need to teach students by way of stan-
dard, proven concepts, and methods. The very power and applicability of these
concepts and methods ensures that they can be understood by everyone and passed
on. Thus, in the rational, scientific world, Newton's concept of mass, Marx's con-
cept of contradiction, or Rawles' concept of justice are univocal and precise. But,
as noted above, this means that to educate our students we impose these concepts
and practices upon them, consequently also forcing their experiences and actions,
that is their worlds, into preexisting, standard concepts, which, after all, are not
politically innocent.
The same is true of allleamers' problems. The learner wants to know how to use
GIS. To proceed with our expertise, which presumably is why we are valuable and
have been brought onto the scene, we translate the learner's vague needs and gen-
eral wants into precise terms. We supply or develop, and then apply, instruments
that will give exact and irrefutable results and indications for practical procedures.
We develop lesson plans that will be maximally functional, that fit with the corre-
late needs of the group and within the prescribed social-economic, aesthetic
norms.
In these cases we exercise our power and accomplish things in the world pre-
cisely insofar as we get the learners to participate in and, thus, continue the preex-
isting and dominant system. To some extent this is good and unavoidable: learners
want and need to leam GIS to become part of the powerful, dominating world.
But, at the same time the result is oppressive to them and ensnares them in a cycle
that continues the processes of oppression (with them now appropriated to con-
tinue what they have intemalized). We know that we also need ways to respond to
the worlds of individuals and groups so that what we come to understand and do
together is generated out of the existential reality of these life-worlds. We are re-
sponsible for not stamping out their specific ways of being in the name of profit-
able and expedient homogeneity. We are responsible for developing ways to see,
attune ourselves to, and nurture the life-worlds of others, including those who
place themselves or are placed in the trajectory of our influence.
Freire agues that the primary way to do this is by disciplining ourselves so that
we can listen to what others have to say and by changing our professional mission
to helping others to say what they want to say in their own terms. In Freire's view,
this means starting with the admission that we do not know what the other per-
son's world is like, nor what their real problems and needs are, much less what
acceptable solutions would be. Nor, likely, does the other person. Ifthey did, they
would not need or consult uso The vibrant relation between learner and teacher is
generated insofar as teachers can help learners to name and become conscious of
their worlds, their needs and possibilities. The process of helping learners to ar-
ticulate their worlds in their own terms is a process of liberation and empower-
ment, for them and for the experts too.
Freire's approach integrates educational, political, and social theory with per-
sonal experience. He contends that the freeing transformation of praxis is achieved
through dialogue in a process (in his words, conscientizacao - conscientization)
Qualitative GIS 333
that allows us to critically assess and und erstand society and our situation in it.
The process begins with investigations that uncover what Freire calls generative
themes, that is, the controlling postulates that are existentially and emotionally
powerful to a group. These themes are then presented back to the group through a
series of often-pictorial codifications in which the teacher elicits distinctions such
as those between cultural and natural dimensions or relationships among inside
and outside groups contending for power.
In this format, where problems are raised for people to discuss in their own
terms, contradictions naturally are discovered; in turn, these can be codified and
presented for further reflection. Thus, the educator can pose a problem to the
group; through dialogue the group begins to surmount the initiallimitations of the
situation. Obviously, the only way for the project to work successfully is for the
participants to engage in genuine dialogue together, for intensive and long periods
of time. Together, and scrupulously avoiding thoughtlessly accepted concepts,
what matters has to be allowed to be named and thought in its own terms, that is,
in terms of the character of each thing and the webs of relationships among them.
In the process, the learners can discover for themselves the contradictions among
elements and systems of meaning, intent, and practice. They can begin to explore
how the contradictions might be overcome in ways that allow the maximum nur-
ture of their world as it discloses itself to them.
To have a more concrete sense of what this means, think of the fieldwork in-
volved in understanding a given geographical realm. We know that it is easy to do
research in the relevant literature, draw out the necessary concepts, devise a hy-
pothesis, and formulate a questionnaire. After a pilot project or two, we are ready
to go, to translate the not-yet-known into the known. But we also all know how we
falsif}r the worlds we are studying when we do so--at least by leaving out so much,
and I would agree with Freire, by violently translating everything into foreign,
standard categories. To remedy this, it is increasingly common to try to go open-
mindedly and see what is there. Then, from initial field observations and conversa-
tions, we devise open-ended interviews, and if that information is not precise
enough, formulate questionnaires. But, these procedures have to do with us com-
ing to know their world. Freire's point is that the opposite needs to happen: the
others need to articulate-delineate their own world, in their own terms. Thus,
though we can assist in the process with our expertise and technologies, our first
obligation is to facilitate visualization and dialogue among the participants, who
thereby articulate their world for themselves and us, as they explain it to uso
Comparative theologian lohn S. Dunne develops a very useful strategy that may
help us to 'pass over' from ourselves to others, and then to pass back (1967).
Dunne begins his reflections with a personal search for what so me take to be the
issues that matter most. 'How can I deal with my fear of my death?' 'Is there a
God?' 'Am I all alone in facing life's difficulties?' These timeless questions have
been encountered by many over the past centuries, but still are mine right now, to
be answered by me, unavoidably. Though each of us has to answer such questions
for ourselves, since others have asked these questions before us (and perhaps even
found 'answers' or at least comforting resting places along the way), Dunne ex-
plores our issue of personal and shared understanding.
334 R. Mugerauer
. . . can be broadened and followed in a much wider context than they ordinarily
would be. The passing over and back, then, tends to bridge the gap between private
knowledge and public knowledge and to give the seeking and finding that occurs on a
strictly individual level something of the communicability of public knowledge.
[Comparing one's personal questions and findings with those of others allows us to be]
... able to pass from the standpoint of our lives to those of others, to enter into a sym-
pathetic understanding of them, to find resonances between their lives and our own,
and to come back once again, enriched, to our own standpoint. (Dunne 1967, viii-ix)
GIS admirably suits itself to such a process. It can provide the means to graphi-
cally present the mapping of one's own world in most whatever way one wishes.
(Remember that built into the very code systems and protocols there are deeper,
fixed limitations that ultimately need to be overcome or removed.) What matters
in a mapping, what is included and excluded (such as the relations among ele-
ments, the means and forms of graphie presentations) would be worked out in
each original application of a system to a newly delineated and articulated world.
Importantly, the decisions that stern openly and responsibly from implicit and
explicit value systems can be respected and built-in from the start. What was not
self-consciously used can come to group consciousness so that its future impor-
tance may be decided. And, since learners would have to visualize-articulate their
own world in their own way and then format that into GIS, they would start with
their own world, pass over into the dominant one, and then back into their own
(now bi-cultural realrn). The teacher would begin in the dominant technological
world of GIS (at least for purposes of the technical facilitation, but not necessar-
Qualitative GIS 335
ily), pass over, at least a bit, in dialogue and work to the world ofthe learners, and
then pass back into her own.
Theoretically and practically, we are justified in holding that 'there is no abso-
lute standpoint, since no standpoint would exhaust the truth of human culture and
built reality, though there is the possibility of our passing over from one contex-
tual horizon to another' (Dunne 1967, 5). Parallel with this, there is no purely rela-
tive standpoint, since though humans operate within specific traditions, disci-
plines, and cultural contexts, one's deep questions, patterns of thought and action,
and way of life do connect with those in other traditions, disciplines, cultures, and
times.
Boundaries need to be acknowledged and respected. By letting boundaries be,
we mark or even celebrate the differences, but are not isolated by them. Crossing
boundaries, then, is not a matter of scientific method achieving objectified knowl-
edge; nor is it idiosyncratic voyeurism. Crossing over and back is possible be-
cause we face not the problem of the unintelligibility of the other, but the inex-
haustible intelligibility of other people, practices, processes, GIS, and other
information technology projects yet to come.
In contrast to the positivistic mental mapping procedures of the dominant GIS
paradigm (recall the quotation above from Gould and White), examples of self-
articulation exist that can be amplified. On the one hand, there are the many grass
roots electronic communities that could implement GIS in the same spirit in which
they now do operate electronically. We all have our favorite community Web
sites. 7 Groups of specialists and ordinary users alike need to collate and share
sources so that we all can Jeam from the entire set whose productions we value.
In addition, to focus on the basic operation of mapping, it would be interesting
and fruitful to transfer to GIS the grass-roots mapping processes underway around
the world, such as documented in Boundaries of Home: Mapping for Loeal Em-
powerment by Doug Aberley et al. (1993). In contrast to the criteria of good = true
= correct in positive science, Aberley contends that
It is important to repeat over and over that there is no 'good' mapping or 'bad' mapping.
Leave the need for perfection to the scientists; what you are being encouraged to do is
honestly dcscribe what you already know about where you live in a manner that adds
momentum to positive forces of change .... every region has the potential to be repre-
sented by as many unique interpretations as it has citizens. Reinhabitants will not only
1eam to put maps on paper, maps will also be sung, chanted, stitched and woven, told
in stories, and danced across fire-lit skies. (1993, 5)
gional information would be added. 8 This is related to my own work with Barbara
Parmenter and an interdisciplinary team of graduate students to generate a Quali-
tative GIS for a neighborhood planning project outside Austin, Texas. The resi-
dents in the Spicewood Corridor, off the Old Spieewoods Springs Road west of
the city, are seeking a way to explore their own identity and that of their local
place in order to begin to imagine ways to develop and keep safe the qualitatively
distinctive environment in which they have chosen to live. This is a still-emerging
version of a conservative Qualitative GIS, in which we are encoding information
about the experiences of the natural environment and personalized individual and
group information onto the standardized GIS databases. 9
A more difficult and yet promising project would be to use basic GIS formats to
generate customized combinations of not-necessarily-representational 'designs'
and 'words' to originarily let a worldview emerge and be named in its own terms.
There is no reason at all why a combination of Freire's proven pedagogy that com-
bines visual representation and naming-dialogue in local, dialectical words cannot
be transformed into visualization that presents other quantitative information and
qualitative interpretations in a democratic, pluralistic GIS system.
References
Aberley, D., et al. 1993. Boundaries 01 Horne: Mapping lor Local Empowerment. Gabriola
Island BC. and Philadelphia PA: New Society Publishers.
Aberley, D. and Lewis, G.M. 1998. Cartographic Encounters: Perspectives on Native
American Mapmaking and Map Use. Chicago: University ofChicago Press.
Albrecht, J. and Lim, G-C. 1986. A search for alternative planning theory: Use of critical
theory. Journal 01Architecture-Planning Research 3: 117-3l.
Amheim,. 1986. Dynamics 0/ Architectural Form. Cambridge: Harvard University Press.
Blackburn, R. 1989. Defending the myths: The ideology of bourgeois social science. In
Cockburn, A. and Blackburn, R. (eds.) A BrielGuide to Bourgeois Ideology in Student
Power. New Left Review, 154-69.
Bravo, M.T. 1996. The accuracy of ethnoscience: A study ofInuit cartography and cross-
cultural commensurability. Manchester Papers in Social Anthropology, no. 2. Manches-
ter: Department ofSocial Anthropology, University ofManchester, 1-36.
Brody, H. 1988. Maps and Dreams. Vancouver: Douglas and Mclntyre.
Brody, H. 1989. Maps and journeys. In Macleod, F. (ed.) Togail Tir Marking Time: The
Map olthe Western Isles. Stornoway, Scotland: Aeair Ud. and Lanntair Gallery, 133-36
Castells, M., Mollenkkopf, R., and Robson, M. 1998. The Rise of the Network Society.
Paper presented at Telecommunications and the City Conference, University of Georgia
(March).
Chatwin, B. 1987. The Songlines. New York: Penguin.
Collins, D. 1977. Paulo Freire: His Life, Works, and Thought. New York: Paulist Press.
8 Doug Aberley, Beatrice Briggs, Kai Snyder, Jonathan Doig, and George Tukel supply
examples of successful case studies and techniques.
9 The Web site for this project is <http://mather.ar.utexas.edu/students/cadlab/spicewood/>
(For information on the Spicewoods Springs Road Project: drbob@mail.utexas.edu).
Qualitative GIS 337
Conclusion
20 From Sustainable Transportation to
Sustainable Accessibility: Can We Avoid a
New Tragedy ofthe Commons?
Helen Couclelis
Department of Geography, University of California, Santa Barbara CA 93106-4060, USA.
Email: cook@geog.ucsb.edu
20.1 Introduction
the case is due to the fact that profound reorganizations of activity patterns are
taking place at aB scales such that the net number of interactions that involve
physical movement rather than electronic contact appears to be increasing rather
than decreasing. As this paper argues, a large part of this phenomenon is due to
the fragmentation 0/ activity that is taking place such that activities that used to be
associated with a single location (e.g., my workplace) are now increasingly scat-
tered among geographically distant locations (e.g., my office, horne, associate's
horne, hotel room, car, train, or plane). Thus the contact set of individuals, the
number of places they interact with, explodes from one location per activity to a
potentially indefinite number of locations. While many of the corresponding inter-
actions are carried out in virtual space, many others are very much physical, plac-
ing old-fashioned demands on transportation systems to meet old-fashioned
accessibility needs.
With more people traveling more often than ever before, sustainable transporta-
tion is emerging as a major theme in contemporary transportation research. The
notion of sustainable transportation is closely associated with mobility, accessibil-
ity, urban form and function, environmental quality, and social and economic life.
Of these, accessibility may be seen as the most central concept, linking the others
together. Hodge (1995) argues that the recent shift in emphasis from mobility to
accessibility in transportation research may not be quite enough to ensure sustain-
ability. The reason is that in the information age the demand for contacts in gen-
eral and especially of the physical kind may be increasing faster than is necessary
for a weil functioning economy, society, and culture: striving to meet without fur-
ther ado the associated proliferating accessibility demands may not be in the best
long-term interest of our cities, the environment, or society. Thus we may speak of
'sustainable accessibility' as the goal of meeting reasonable mobility and accessi-
bility requirements of individuals while reducing the need for ever larger contact
sets now and in the future.
This paper argues that sustainable accessibility is a precondition for sustainable
transportation, and explores the meaning of that phrase within the context of rap-
idly spreading information and communication technologies (ICTs ). It begins by
highlighting the complexity of the notion of accessibility and its relation with dif-
ferent views of transportation technology. It then discusses how accessibility may
be redefined in the information age, in pace with other momentous changes in how
places, activities, and spatial relations are being conceptualized within a new hy-
brid space-time that is partly traditionally geographic, partly virtual. Central to
these changes is the fragmentation of activity brought about by ICTs. That obser-
vation leads to an exploration of the notion of sustainable accessibility within the
framework of the 'tragedy of the commons' dilemma. The paper closes with a
brief discussion of some of the many research challenges, both theoretical and
practical, lying ahead in this brave new domain.
From Sustainable Transportation to Sustainable Accessibility 343
Underlying virtually all conventional theories and models of urban land-use struc-
ture and planning is the assumption that an activity is associated with one place at
one time: work with a workplace, shopping with a commerciallocation, schooling
with an educational facility, and so on. Tell me where you are, I can tell you what
you are doing. What is more, tell me where you work and I may be able to tell you
where you are likely to live, because I know that you are trying to optimize acces-
sibility to your most important destinations. In the information age by contrast, to
the extent that an activity involves exchange of information between or among
people, or between people and machines, that activity may be fragmented into
tasks that are widely distributed over space and across time. Whatever activity
they may be engaged in: work, recreation, shopping, education, people are in-
creasingly likely to rely on a variety of physical and digital means to gain access
to the people, tools and information they need. Some ofthese contacts can only be
realized through physical movement, some can only be realized electronically, and
some others may be realized in either way. Each of these contacts involves a geo-
graphic location where the contacted person or device is, so that several locations
may have to be accessed by the appropriate combination of travel and telecoms for
a single activity to be carried out. It is this interweaving and mutual dependence of
physical mobility and electronic communication, not merely the spreading use of
ICTs, that defines the major information-age challenge for accessibility research.
Because of these changes the familiar one-to one mapping: activity => place
becomes a one-to-many mapping: activity => places, or even a many-to-many
mapping: activities => places. The same is true of time, as the proverbial nine-
to-five weekday job gradually gets fragmented into chunks spread out over arbi-
trary hours of the day (and many of the night), interspersed with tasks from other
activities occurring at equally odd - by traditional standards - times as well as
places. The colonization of time, as this phenomenon has been called, does not
concern only the business executive who is talking with London or Tokyo in the
middle of the night just as the financial markets start trading in distant time zones.
The secretaries and the janitors, the computer support personnei, the food service
and transportation workers, the gym and the grocery store and the plumbers and
baby sitters are all part of the same nexus of interdependent activities that drive
From Sustainable Transportation to Sustainable Aeeessibility 347
Table 20.3. A taxonomy of possible spaee and time distributions of aetivities in the infor-
mation age
Loeation of aetivity Time of aetivity
all at one loeation all at one time
at alternate loeations at alternate times
distributed along a route in time sequenee
distributed aeross spaee at several different times
ubiquitous (anywhere) any time
nowhere never
Hence the central thesis ofthis paper: It is not distance that is dead; it is activity
that is disintegrating. Activity becomes a distributed space-time process, a net-
work of material movements and digital contacts, an interweaving of (electronic)
communication and (physical) action. We thus rejoin the earlier mentioned discus-
sions of place as process, communication, or network. Understood from the func-
tional, socioeconomic, cultural, rather than the strictly spatial point of view (see
Table 20.1), these outlandish-sounding notions of place amplify the idea of the
fragmentation of activity, and raise some difficult issues regarding the meaning of
territory, community, and identity in the information age (Curry and Eagles,
1999). The purview ofthis paper is however much more modest, as it only focuses
on the spatio-temporal and functional dimensions of the restructuring of place and
activity, and on the implications for accessibility of that fluid but still somewhat
tangible context. From this restricted perspective I propose the following as a
working definition of accessibility in the information age: the ability to access,
either physically or electronically, and at the appropriate time(s), all the locations
that are necessary or desirable for participating in a given activity. This implies
that there are three kinds of knowledge necessary for the study of accessibility: (1)
the distribution in physical space-time, relative to actual or potential activity par-
ticipants, of the loci of the component tasks constituting specific activities; (2) the
structure and characteristics of the access-enabling technologies, both physical and
electronic; and (3) the relations of substitution, complementarity, and synergism
among available physical and electronic options (Abler and Falk 1981). The terms
space-acijusting technologies (Abler 1975, Janelle 1991) or spatial technologies
(Couclelis 1994), designating the complex of information, communication and
transportation technologies that together work to modify spatial relations, serve to
underline the need for a common approach to two materially very different but
functionally very similar approaches to overcoming spatial separation.
Figure 20.1. Physical and electronic contacts and the possibility for substitution
dependent contact demand. The first two components of the hypothesis, which are
already weil supported by data, can be expressed in simple notation and illustrated
with a Venn diagram in Figure 20.1.
Can, and should transportation keep up with these ever-increasing demands? In
recent years the traditional goal of transportation planning has shifted from an
almost exclusive emphasis on facilitating mobility to a greater concern for acces-
sibility, in recognition of the fact that travel is (for the most part) a means towards
the end of accessing a desired or needed destination. Mobility demands are thus
better seen in the context of more fundamental accessibility requirements, which
in turn are closely linked with spatio-temporal patterns of activities at both the
individual and the systemic scales. While the goal of facilitating accessibility may
seem beyond reproach, some researchers doubt that striving to meet without fur-
ther ado the ever-mounting accessibility demands of the post-industrial or infor-
mation society is a viable strategy for transportation planning (Hodge 1995). One
may argue instead that the goal of sustainable transportation should be to meet
reasonable accessibility demands, that is, to facilitate mobility where it is really
necessary. Mobility is not really necessary where there are good alternatives to it,
as they are in the case where appropriate ICTs can substitute for physical move-
ment. Referring to the diagram above, there are two ways to slow down the
growth of the demand for physical contacts, (C p): increase the possibility for
travel substitution, or increase the attractiveness (relative utility) of substitution.
Both aim at enlarging the overlap area (Co) on the diagram: The first will ensure
that people have a choice; the second will increase the likelihood of people mak-
350 H. Couclelis
ing the correct choice from the viewpoint of sustainable transportation (Janelle
1995). This leads to the se co nd hypothesis:
Hypothesis 2: The utility of travel remains too high relative to its alternatives.
People like to travel not just because they still tend to value face-to-face contacts
and real-world destinations over their virtual counterparts but also because, con-
trary to what traditional transportation research has always assumed, travel itself
has considerable intrinsic utility. Modem automobiles are increasingly comfort-
able, pleasant and safe moving 'places' where people enjoy spending part of their
day. Inside our cars we entertain and educate ourselves, work and daydream,
smoke and snack, watch our kids or socialize with friends, nurture our self-esteem
as smart and skilled drivers. Style, pleasure, comfort and luxury are increasingly
becoming selling points in public transportation as well, at least for those who can
afford the first-class ticket. ITCs save time and money but have yet to meet the
psychic utility of driving a late-modelluxury car or even of sharing a bus seat with
a favorite fellow commuter. Undoubtedly the modal choice problem within the
growing set of spatial technologies is a much more complex one than that of
choosing among physical transportation alternatives. Fischer, Ricco, and Rammer
(1992) and Fischer and Rammer (1992) have investigated these kinds of choices
involving both physical and electronic options within contact networks of academ-
ics. It is clear that the utilities leading to such choices are multidimensional and
are influenced by a variety of characteristics pertaining to the initiator of the con-
tact, the contacted destination, the costs and benefits of the available options, and
the familiarity ofthe contact initiator with these options and their availability. This
is not much different from what is known from numerous studies of discrete
choice behavior in general and traditional modal choice research in transportation
in particular. These studies demonstrate that well-established methods of analysis
can in principle be extended to the study of information-age behaviors within the
ever-expanding set of spatial technologies. The highly interdependent choices of
contact mode made in weaving together component tasks of fragmented activities
will of course pose novel, additional challenges. These are likely to be practical
much more than conceptual, as these micro-Ievel approaches appear well suited in
principle for studying how individuals cope with these more complex choice situa-
tions.
offs is such that while the collective optimum is reached when both players opt to
cooperate, individual rationality in ignorance of the other's choice dictates defec-
tion. More specificaIly, a prisoner's dilemma situation occurs whenever the struc-
ture of payoffs obeys the following ordering, regardless of the actual numerical
values ofthe terms (Hardin 1982):
T>R>P>S
where
R = reward for cooperating (collective optimum)
T = temptation to defect for the individual decision maker
=
P punishment for not cooperating (collective cost)
=
S 'sucker's payoff' for being the only nice person
The prisoner's dilemma and its various extensions, especially the iterated and n-
person versions, have been used widely in social choice research to investigate
individual cooperation and defection and the ensuing generation of negative exter-
nalities in a collectivity. Real-world examples abound: a water or power system
collapses during aperiod of peak use, despite dire warnings, because people are
not prepared to inconvenience themselves by cutting down on their own consump-
tion if they believe that most others will not. In a similar vein, in the absence of
external coercion, few car buyers will be persuaded to buy the less glamorous but
less polluting model, few fishermen to limit their harvest so as not to deplete the
fisheries, few developers to spare the area they live in from over-development, if
they know or believe that most others will not. These are all cases where individu-
als have a choice to cooperate, at some personal cost, towards some common goal,
or to do what is best for them individually - best, that is, as long as all or most of
the others do not decide to do the same thing. The PD-like structure of payoffs is
such that even weIl meaning, not necessarily selfish individuals are deterred from
attempting to cooperate by the belief that their small sacrifice will be in vain.
These kinds of phenomena often arise in a spatial context, as several geographers
and planners have found out (see, for example, Herniter and Wolpert 1967, Od-
land 1985, Couclelis 1989).
The case of spatial technology choices discussed in this paper is analogous to a
particularly common version of the n-person PD known as the tragedy 01 the
commons (Hardin 1968). This describes the depletion or degradation of a common
property resource or public good to which people have free and unmanaged ac-
cess. Here the common resource threatened by degradation is the transportation
system in its environmental and socioeconomic context (its 'sustainability'), and
the binary choice for individuals is whether to achieve a desired contact through a
physical trip (delect) or ICTs (cooperate), assuming that the option is available.
The other terms of the payoff structure are interpreted as follows:
352 H. Couclelis
The well-known work ofAxelrod (1984, 1997) and others on the iterated PD
tempers the cynicism of the original PD by showing that cooperation makes good
sense for individuals involved in repeated transactions with others. Also,
Schelling's (1978) cIassic analysis of the n-person version demonstrates that a
collective optimum is often possible even when (or provided that) some fraction of
the participants opt to defect. Schelling's work has especially interesting implica-
tions for policy making because it shows that in the n-person case several different
outcomes are likely depending on the relative numerical values of the payoffs,
rather than just their rank orderings. In some cases manipulating slightly the val-
ues of individual rewards and punishments (in this case: the utility of telecontacts
versus the utility of physical travel) can help steer the system towards its collec-
tively optimal state despite a number of 'free riders' who will take advantage of
others' restrain. Indeed, the optimal transportation conditions are not achieved
when no one is traveling but when just enough are traveling at any one time so as
to keep both the trafiic flowing and the gas stations in business. The value of this
kind of analysis lies in the possibility to quantify the obvious qualitative fact that
deliberate policy measures that make travel less attractive or that increase the util-
ity of telecontacts are likely to encourage ICT substitution for traveI. On the other
hand, the analysis also suggests that changing relative utilities and feedbacks
within a dynamic system are Iikely to bring about adjustments towards the collec-
tive optimum even in the absence of any external intervention. In a reasonably
rational world, as more and more people are deterred by the individual costs and
negative externalities of increasing congestion (decreasing T, increasing P), more
and more should be seeking out alternatives. The self-correcting property of high
congestion has often been argued but less often observed on our transportation
networks. It is conceivable that in the not too distant future ICTs will offer the
higher-utility alternatives that other conventional measures of trip reduction have
not been able to provide.
The implications of the notion of sustainable accessibility range weIl beyond
transportation systems to urban structure, the environment, the economy, and so-
ciety. Conventional planning wisdom holds that compact rather than diffuse urban
forms, mixed rather than segregated land uses, reliance on public rather than pri-
vate transport -for example, European-type rather than American-type cities -
help achieve higher levels of overall accessibility with lower negative socioeco-
nomic and environmental impacts. This may no longer hold true in the informa-
tion age. Indeed, it is likely that in settings with high overall physical accessibility
there is less incentive for individuals to meet the proliferating demands for con-
tacts through virtual rather than physical interaction, leading to increasing pres-
sures on the transportation infrastructure with all the associated societal costs. We
From Sustainable Transportation to Sustainable Accessibility 353
have seen something like this happen time and time again in the best laid out
modem cities where neither accessibility nor transportation proved sustainable in
the long run. What is new is that planners now have even less control over the
location of today's fragmented activities (and the ensuing growing transportation
demands ) than they did when living, working, shopping, leaming, or recreating
were each to be found in their own pre-assigned places. Is it a feasible goal for
transportation to make everywhere easily accessible from everywhere? Are acces-
sibility and sustainability public goods vulnerable to the 'tragedy of the com-
mons'? If so, what societal and planning responses are appropriate? The research
challenges are formidable, and we have just begun to scratch the surface.
This paper has made several claims that demand to be tested through both empiri-
cal research and more rigorous theoretical development. These are as folIows:
First, that a useful conceptualization of accessibility in the information age may be
based not on the 'death of distance' metaphor but on a new functional conception
of activity (and perhaps also place) as spatially fragmented. Second, that the frag-
mentation of activity directly contributes to the somewhat paradoxical rapid
growth in travel demand especially noticeable in countries with the highest rates
of ICT development. Third, that because of the growing spatio-temporal plasticity
and fragmentation of activities, planners have less control than ever before on
what activities take place where (and when). Fourth, that planners should perhaps
not strive to keep up uncritically with pro li fe rating mobility demands but espouse
instead the notion of sustainable accessibility. Fifth, that physical travel remains
too attractive relative to viable ICT-based alternatives, but that the 'tragedy of the
commons' some fear will result may turn out to be to so me degree self-correcting.
To what extent these sweeping statements may be true or helpful for understand-
ing accessibility in the information age, or for informing planning and policy mak-
ing in related domains, cannot be said for sure at this point. As the chapters in this
book have shown there are too many defining aspects of the issue where either the
data or the appropriate conceptual frameworks are simply lacking. Concerning the
notion of sustainable accessibility highlighted in this essay, perhaps the most criti-
cal empirical question is finding out how the relative growth of physical and digi-
tal contacts in people's contact sets may be affected by a range of factors that may
differ widely from place to place. This chapter is based on the U.S. experience and
may not reflect how the issues discussed may be perceived from the viewpoints of
other nations. Comparative studies are very appropriate here, examining, for ex-
ample, regions of the world that are similar in general socio-economic terms but
different in terms of some obviously important variables such as spatial structure,
ICT presence, or degree of planning contro!. Such research could look at the gen-
erally different structures of ~ say ~ European and American cities (e.g., high-
density vs. low-density, public-transit oriented versus automobile oriented); the
different rates of Internet connectivity, access and use in Europe and the United
354 H. Couclelis
Acknow ledgements
This paper was first presented at the conference on Social Change and Sustainable Trans-
port (SCAST), co-sponsored by the U.S. National Science Foundation (NSF) and the Euro-
pean Science Foundation (ESF). My thanks to Professor Bill Black and the other SCAST
conference organizers for their generous support, and to Professor Hugo Priemus and his
group at the TU Delfl, the Netherlands, for thoughtful comments and another opportunity to
air these ideas.
References
Abler, R 1975. Effects of space adjusting technologies on the human geography of the
future. In Abler R, Janelle, D., Philbrick, A. and Sommer, J. (eds.) Human Geography in
a Shrinking World. North Scituate MA: Duxbury Press, 35-56.
Abler, RF., and Falk, T. 1981. Public information services and the changing role of dis-
tance in human affairs. Economic Geography, 57: 10-22.
Adams, P. 1998. Network topologies and virtual place. Annals ofthe Association ofAmeri-
can Geographers, 88: 88-106.
Axelrod, R 1984. The Evolution ofEooperation. New York: Basic Books.
Axelrod, R. 1997. The Complexity ofCooperation: Agent-based Models ofCompetition and
Collaboration. Princeton Studies in Complexity. Princeton NJ: Princeton University
Press.
Bolton, R. 1997. 'Place' as 'network': Applications ofnetwork theory to local communities.
Volume in honor of T.R. Lakshmanan, Department of Economics, Williams College,
Williamstown MA 01267, December.
Bureau of Transportation Statistics (1997). American Travel Survey 1995, BTS/ATS95-US.
US Department ofTransportation, Washington DC.
Cairncross, F. 1997. The Death of Distance: How the Communications Revolution will
Change Dur Lives. Boston MA: Harvard Business School Press.
Castells, M. 1989. The Informational City: Information Technology, Economic Restructur-
ing, and the Urban-regional Process. Oxford: Blackwell.
From Sustainable Transportation to Sustainable Accessibility 355
Castells, M. 1996. The Rise of the Network Society. The Information Age: Economy, Soci-
ety, and Culture, 1. Oxford: Blackwell.
Castells, M. 1998. End of Millennium. The Information Age: Economy, Society and Cu 1-
ture, 3. Oxford: Blackwell.
Coucielis, H. 1989. Macrostructure and microbehavior in a metropolitan area. Environment
and Planning B 16:141-54.
Coucielis, H. 1994. Spatial technologies. Environment and Planning B: Planning and De-
sign 21:142-43.
Coucielis, H. 1996a. The death of distance. Environment and Planning B: Planning and
Design 23:387-89.
Coucielis, H. 1996b (ed.). Spatial Technologies, Geographic Information, and the City., TR
96-7. National Center for Geographie Information and Analysis, Santa Barbara CA.
Coucielis, H. 1998. The new field workers. Environment and Planning B: Planning and
Design 25:321-23.
Curry, M., and Eagles, M. 1999 (eds). Place and Identity in an Age of Technologically
Regulated Movement, National Center for Geographie Information and Analysis, Santa
Barbara CA.
Fischer, M.F., Ricco, M. and Rammer, C. 1992. Telecommunication media choice behavior
in academia: an Austrian-Swiss comparison. Geographical Analysis 24:1-15.
Fischer, M.F. and Rammer, C. 1992. Kommunikationsnetze von Wissenschaftlern: Ergeb-
nisse einer Fallstudie and Wiener Universitaeten. Mitteilungen Der Oesterreichischen
Geographischen Gesellschaft 134: 159-76.
Graham, S. and Marvin, S. 1996. Telecommunications and the City: Electronic Spaces,
Urban Places. London: Routledge.
Hardin, G. 1968. The tragedy ofthe commons. Science 162:1243-48.
Hardin, R. 1982. Collective action as an agreeable n-prisoner's dilemma. In Hardin, R., and
Barry, B. (eds.) Rational Man, Irrational Society. Beverly Hills: Sage, 123-35.
Herniter, J. and Wolpert J. 1967. Coalition structures in the three-person non-zero-sum
game. Peace Research Society: Papers 8:97-108.
Hodge, D.C. 1995. Intelligent transportation systems, land use, and sustainable transporta-
tion. Paper presented at the ITS America Alternative Futures Symposium on Transporta-
tion, Technology, and Society, Washington DC.
Janelle, D.G. 1991. Global interdependence and its consequences. In Brunn, S.D. and Lein-
bach, T.R. (eds.) Collapsing Space and Time: Geographic Aspects ofCommunications
and Information. London: Harper Collins Academic, 49-81.
Janelle, D.G. 1995. Metropolitan expansion, telecommuting, and transportation. In Hanson,
S. (ed.) The Geography ofUrban Transportation. 2d ed. New York: Guilford Press, 407-
35.
Johnston, R.J. 1997. The end of distance but a continuing bounded journey through space?
Environment and Planning B: Planning and Design, 24:319-22.
Knoke, W. 1996. Bold New World: The Essential Road Map to the Twenty-first Century.
New York: Kodansha International.
MitchelI, W.J. 1995. City of Bits: Space, Place, and the Infobahn. Cambridge MA: The
MIT Press.
Odland, J. 1985. Interdependence and deterioration in the housing stock of an American
city. In Haynes, K. E., Kuklinski, A., and Kultalahti, O. (eds.) Pathologies ofUrban Pro-
cesses. Tampore: Finn Publishers, 323-39.
356 H. Couclelis
5.1 An exploded view ofthree binary masks for a rectangular area over
five time periods 77
5.2 Extending time-geographie eoncepts 80
5.3 Potential aeeess and interaction in Auekland 81
5.4 Potential impacts of physical/virtual substitution 86
12.1 Telephone availability for various countries vs. the speed ofthe Internet
connection (measured using latency) between those countries
and the Uni ted Kingdom 206
12.2 Gross Domestic Product (GDP) for various countries versus speed
ofthe Internet connections (measured using latency) between those
countries and the United Kingdom 206
Figures 359
12.3 Latency distance decay curve derived in the Murnion and Healey (1998) study 208
12.4 Locating a remote object 0 using triangulation 209
12.5 Correlation on ten routes between latency in each direction plotted against
the average latency in both directions for that route 211
12.6 The average latency measured in one direction plotted against the average
latency measured in the reverse direction for 10 separate information flow
routes 211
12.7 The locations ofthe computers used in the Internet triangulation exercise 212
12.8 SimuItaneous latency measurements taken for the attempted triangulation 212
12.9 Latency predicted by the neural network against the actual measured latency 213
12.10 Latency between Helsinki and Cape Town 214
14.1 A two-dimensional representation ofthe three map layers after transformation 250
14.2 A multi-scale, 3D representation ofthe individual's space-time paths 251
14.3 An extensibility diagram ofa set ofhypothetical activities 252
20.1 Physical and electronic contacts and the possibility for substitution 349
Tables
DAVID C. HODGE is Dean ofthe College of Arts and Sciences at the University
of Washington. His Ph.D. is in Geography trom Pennsylvania State University. He
served previously as Chair of the University of Washington's Geography Depart-
ment, as Program Director for the National Science Foundation's Geography and
Regional Science Program, and as Editor of The Professional Geographer. Re-
search has contributed to understanding equity issues in urban social geography
and transportation, and to the impact of information and intelligent transport tech-
nologies on the spatial form ofmetropolitan regions.
HARVEY J. MILLER has a Ph.D. in geography trom The Ohio State University
and is currently Associate Professor of Geography at the University of Utah. Dr.
Miller is also North American Editor ofthe International Journal ofGeographical
Information Science. His research and teaching interests focus on Geographic In-
formation Systems for Transportation (GIS-T) and geocomputational methods for
380 Contributors
spatial analysis. He is weil known for his work on modeling accessibility using
space-time prism concepts.
MITCHELL L. MOSS is Director of the Taub Urban Research Center and is the
Henry Hart Rice Professor ofUrban Policy and Planning in the Robert F. Wagner
Graduate School ofPublic Service at New York University. His Ph.D. is in Urban
Studies from the University of Southern California. He has written extensively
about the diffusion of telecommunications technologies, the role of telecommuni-
cations in altering urban landscapes, and on policies regarding economic devel-
opment, information cities, and the global economy.
areas, the use of Exploratory Data Analysis for evaluating spatial data accuracy
within GIS, and most recently, changes in accessibility patterns in the spatial
structure of employment within the metropolitan region of Los Angeles.
QING SHEN is the Mitsui Associate Professor of Urban Planning at the Massa-
chusetts Institute of Technology. His Ph.D. is from the University of California,
Berkeley. His research has concerned the modeling of relationships of urban spa-
tial structure, transportation, and telecommunications. In addition he has investi-
gated the office growth of metropolitan America, and in the impact of
metropolitan restructuring on employment accessibility and central cities.
H.von Holst, A. Nygren, R. Thord (Eds.) H. von Holst, A. Nygren, A.E. Andersson
Transportation, Traffic Safety and (Eds.)
Health -The New Mobility Transportation, Traffic Safety and
First International (onference, Göteborg, Health - Man and Machine
Sweden, 1995 The contributions of th is volume
This collection of contributions by demonstrate the importance of a multi-
scientists and adm inistrators presents disciplinary approach in an effort to
some of the latest findings in the area of improve various prevention programs
mobility and its relation to medicaJ tre· aiming at reducing mortality, morbidity,
atment, rehabilitation, public health and disability and handicap in the traffie.
prevention. 2000. x, 190 pp. 27 figs" 23 tabs.
I 997. X. 226 pp. Hardcover' DM 128,-
Hardcover' DM 142,- ISBN 3-540-67443·8
ISBN 3-540-62524-0
H. von Holst, A. Nygren, A.E. Andersson
(Eds.)
H, von Holst, A. Nygren, A.E. Andersson
(Eds.)
Transportation, Traffic Safety and
Health - Human Behavior
Transportation, Traffic Safety and The development of modern vehicles
Health - Prevention and Health and new traffic systems requires more
The book gives a broad and deep over- sophislicated behavior and technology.
view of the strategies that have to be New medical technologies such as
adopled 10 avoid the rapid inerease in improved neuropsychologie methods
road trafflc injuries. and descriptive mapping of behavior
2000. X, 194 pp. 25 flg •., 21 tab •. with imaging techniques facilitate the
Hardcover' DM 128,- understanding of the anatomy and phy·
ISBN 3-540-67444-6 siology of human behavior.
2000. XIV, 423 pp. 95 fig s., 45 tabs.
Hardcover ' DM 169,-
ISBN 3-540·67445·4
19991(1.116 . . . . . . . 11 ......
_ ' D M . I ! 1'.501" 1>01Lil.IOUJO l . Hed~..,d, B. Und .tlÖm IEds.)
I!II< J.S~~111-1
---_.-
Papers in Journal 01
.~
+-
The Annals 01
--
Regional Science Geographical Regional Science
The Journal oftllt Systems An InttrMtlonal Joumal of
Regional Sden<e Urban, Regional iilnd
Ge-ograph lul lnformillion,
Assodation InterMt lonal Analysis, ThfOry, and Environmental Rtstar<h
Tbe offLeial journal of thr D«ision and PoliC1
RfIIlional Seienee Anoci.· Qffki,tl Joum~1 of tne
The Journal of
tion lnlunational, encour· Western Regional
Gwgraphicll Systems, a
Iging high·quality scholar.
mip on. broad rangr of
journal dtdicattd 10 gw- Science Association
graphie.l information,
topia in the field of rflllional A qu.rlerly journal;n thf
scitnee. Theu topia in· analysi" thwry.and deci· inttrdisciplinary fjeld o(
dude, bUI.,.., not limittd 10, sion,.ims to encourage and
regional.nd urhan studi ....
promole high·qualilY schol-
bfh."ionl modfling of 10' I!S purpose i.!O promote
.rship on import.nt thw·
calion, tranJportltion, Ind high quo1ity scholarmip On
-
I'fticallnd proctical inun
migration drciJion .. llnd thc importlnt IhMrelic.!
in rflllionil scirner, gc-ogra-
uu and urban d~lopmrnt. and empirieal i.. ur' in
phy, Ihe environmenlal sci-
inlerindu$lry analysis, en,,;- I'fg;onalsdenef,
enee"lIld planning. An
ronmmlal and «ological Sub1eriptiOft in/ouTIo1tiOft
analysis, reSOurce manlge- important goal oflhr jour-
nal iSlo encourage interdis-
ment, urban and rflllionil VoI.~.' Inun DM S98
ciplinlry eommunicalion
poliey Inalysi .. gwsrophical
and I'fst.rch,rp«ially whrn
-
ISSN051G-t8M
-
information sYSlem .. and
sp"lial Jtllistics, 'patial Inaly,i" sp.ti.l Ihf<.>- Tn!<!No. t68
ry .nd splli.1 d«ision sys·
_
Subscription infOfm.ltion 15SN tON1S92
lem' Ire Ihr Ihemrs. (eleclronic fdltionl
For in>l~U1ioNl...,b1cribf... Sub1e,iption InfOfrMtion
---
~" 19. 4 I~,u.es DM 298
Vo/ume Volume 2.4 i,wes DM 298
15SN 1056-8190
TM!<! No. 1011t)
ISSN 1435-5930
TItie No. 10109 ..
. . . . . · C_§oMc.
In_. .
_.-
(elKlfonic fdilion) (e!«t,onic fd~ionJ ' "'-'IIU U1)<"lOO
,,,:[tU J1 )lU·nt
_ _ u.s,,,,,c-_ "'*-IfO
.. _ _
........ t _
.... ~ .., ....
$uI>w\pllon 2000:
Fa< IISN......-. «111 100_'",
.,.,..""iot'"-....
Fa<""""""""IISN..-.....,
_~ ...... p/<ow«WIkKI:
__ ,.0. ... ,""
~SOonot"'
--;",.
90S
, Springer
.........
~_611C11MA
...,. ..".....
_ ......... ""' .... f'ri«i ..., .." ..
_(217)111'-
fo:c lll7) 1l1·J06S OU_'...... Io<oI'·ATi'_';-..
-~
R. Cape11o, P. NiJbmp, G. P. pplng
Sustainable Cities and Energy Policies
Tht lim oflhi. book i.lo highlighllht grell J. SIIII ... ell, 5.Geertman,S. Open.ha ... (Ed s..)
poltnlial of d.unl,..Jiud (i.e.loul 0' ur!»n) Geographicallnformation and
rn.,gy policit. in Ichitving tnvironmentaUy·
t.tnign dtvrlopm.nl' for mOOr,n cili ... Urbln
Planning
.... I.in.bilily i. ploc~ in In. Cont.~1 of Ih. IHl1\Orutrata IM .Ial. of IM .rl in IM u.w of infot·
de!»l' on global .u.tainlbl. dowlopm.nl. Th. nulOon $)'Siems and moddling methods in different
main qUeltion addr~ is: whieh ar' th. critieal sp;iliaUgeographicai planning contuts.ll pn:widtl.
suee", faclors for . ucc.... (ul ur!»n .n.rgy poli· rtvitw of dtvdopmmts in Ihr '990S and detaikd
cies! It is also dult wi1h in I mOla.lnalyli, eon· insighl' into IM currenl Ipplication of go:ographical
lut by au...ing Ind comp.ring Ih. perfOlrmance infornution ItdmoIogy in urbon, phyo;caI.. mviron·
Ol( en.rgy polki.. in va,ioul European eiti ... with mental and socio·«(momM;: planning. Imporllnl
• parTieul.r view to r.nrwabi. en.rgy. advances in the uoc 0( Ihr lnter.m for iK(<'$S 10 do!a
'99'l_11l1ll2P1')Q~J7' _ and \I) GIS for planning a~ abo coruidmd.
_ DM'49"
,99'l .... SOpp.l')fI9!....l .....
I\.BNl·~"
.... "'_DM 169"
IS8Nl·~1·1
MN )-'I4I).6§1I'·)